[Fei Li] 14:02:30 Mayor. She's in the zoom. [Fei Li] 14:02:37 So cool, that's [Fei Li] 14:03:07 We got that time, so… [Fei Li] 14:04:08 I don't actually want to… I'm just halfway back. [Fei Li] 14:04:35 So, I'm careful about personalization. Oh, amazing that TDR… [Fei Li] 14:05:00 For that. Took a year, so we sent messages. [Fei Li] 14:05:13 You don't do want to come. [Fei Li] 14:05:18 A whole… well, we're shown to aware that you need 5… [Fei Li] 14:05:51 Porsche… You need to check out for… Firstly, I understand the… pushes around Oh, not like Kanas Nagavando. [Fei Li] 14:06:02 Like, uh, you sign up Udoka. Good. Hi, y'all. [Fei Li] 14:06:07 Put that in the PDF of the bullshit. [Fei Li] 14:06:20 Oh, there you go. Hmm. [Fei Li] 14:06:46 Formal pedagogy around the changing video vision. I'll follow along. [Fei Li] 14:06:53 You kind of wins the age. We'll go, we'll give you more. [Fei Li] 14:07:08 Listen to eat them. [Fei Li] 14:07:21 I'm kind of sitting at home. [Fei Li] 14:07:34 Hi, all y'all. [Fei Li] 14:07:47 So, you're gonna unlawful, huh? Good. [Fei Li] 14:07:52 Especially there. But at the articulation. [Fei Li] 14:08:17 They actually said that. [Fei Li] 14:08:27 Have fun! [Fei Li] 14:08:33 Take a… take a while. [Fei Li] 14:08:38 You have one. Let me explain, uh… We didn't move into the negative. [Fei Li] 14:08:55 Which one, there's no… [Fei Li] 14:09:01 Traditionally would… yeah. [Fei Li] 14:09:12 Amen. Zoom. Welcome. [Fei Li] 14:09:48 Beat axe. [Fei Li] 14:09:57 The bill… [Fei Li] 14:10:02 Commercial boda. That's a choice, actually. [Fei Li] 14:10:16 Okay, Bamanga. About taxi food. [Fei Li] 14:10:28 No? No, absolutely social edition. [Fei Li] 14:10:40 If doubtful, we're going to hide some. [Fei Li] 14:10:49 Very kind. Which, you just… I think that God will… Legal, can you find out more about? [Fei Li] 14:11:10 She got no fun timing, too. [Fei Li] 14:11:27 Look all again, see this… So, beyond… But who here on the internet, who have one idea bug you. [Fei Li] 14:11:39 November, she… [Fei Li] 14:11:50 It's about who came from an event that transit from a… The only UN function. [Fei Li] 14:12:02 That. [Fei Li] 14:12:26 It's a good employee there. [Fei Li] 14:12:36 It's a siren. The user has the PDF limited. [Fei Li] 14:12:49 Aha! [Fei Li] 14:13:06 Yeah. I took classes. [Fei Li] 14:13:18 Yeah. [Fei Li] 14:13:28 That's just… He may be out of your food. [Fei Li] 14:13:42 It's education. Yeah. [Fei Li] 14:13:52 Period. [Fei Li] 14:14:01 Acrobat? [Fei Li] 14:14:07 Which is… [Fei Li] 14:14:15 Would you kick off the Yiji? I said, hey, I don't… How the… result to the SEO. [Fei Li] 14:14:46 It is a sense of the yard show. Use a toolkit. [Fei Li] 14:14:53 Didn't really talk much energy. [Fei Li] 14:15:13 Chitanza, Tamachalla. [Fei Li] 14:15:23 Okay, Kai. [Fei Li] 14:15:38 The machines, uh… [Fei Li] 14:15:47 Where I… [Qidong Zhou] 14:15:58 No, Farrah. You know, fair question. [Fei Li] 14:15:58 You see the whole council. [Fei Li] 14:16:14 The whole… a soup. [Fei Li] 14:16:18 Use the prototype won't you prototype. System. [Hongyu Zhang] 14:16:18 Depending没有共享屏幕。 [Fei Li] 14:16:30 You know, bullshit. [Fei Li] 14:16:41 Moi. [Qidong Zhou] 14:16:52 To her make a folio, then? [Fei Li] 14:17:09 To some TDRL, [Fei Li] 14:17:22 Chai Xinghuang. [Qidong Zhou] 14:17:29 Uh, nothing at all. [Fei Li] 14:17:29 The ancient sea… Shenza, um… [Hongyu Zhang] 14:17:31 他可以听到吗？ [Qidong Zhou] 14:17:36 Not in Oklahoma. [Fei Li] 14:17:41 What chance of microphone? [Fei Li] 14:17:50 Wait, wait, wait. [Hongyu Zhang] 14:18:04 Oh，我可以听到周启东老师说话。 [Qidong Zhou] 14:18:04 Way, way, way. [Hongyu Zhang] 14:18:09 我在线上。 [Qidong Zhou] 14:18:09 If you're also never teamed up. [Fei Li] 14:18:16 Right? [Hongyu Zhang] 14:18:18 哎，我可以听到。 [Qidong Zhou] 14:18:18 Veh… [Fei Li] 14:18:24 That's for sure. [Fei Li] 14:18:37 I'm pretty sure. [Fei Li] 14:18:43 Yeah, a little troubled alum. Why are the only young men? [Fei Li] 14:18:52 It's okay. Pablo. [Qidong Zhou] 14:18:54 Very, very, very, very, very, very. [Fei Li] 14:18:58 Oh. Wait, wait, wait, wait. Sure, Andrea. [Hongyu Zhang] 14:19:05 可以听到吗？可以听到吗？ [Fei Li] 14:19:19 Central. [Fei Li] 14:19:27 We're in there. [Jingzhou Zhao] 14:21:05 Wait, wait, wait, wait, wait, wait… Right? [Qidong Zhou] 14:21:07 Right? [Qidong Zhou] 14:21:13 Now, your official… [Qidong Zhou] 14:21:21 Okay, then I came out. [Jingzhou Zhao] 14:21:21 Sure. Huh? So, my account… [Qidong Zhou] 14:21:28 You cannot control in Kimodo. [Jingzhou Zhao] 14:21:30 Uh, not half of them. [Jingzhou Zhao] 14:21:39 To know Ana. Money, somehow. [Jingzhou Zhao] 14:21:53 We just shoot him. I managed to call that quantity killed one. [Jingzhou Zhao] 14:22:19 The PDR gene. [Jingzhou Zhao] 14:22:34 You don't… [Jingzhou Zhao] 14:22:50 That's true. She can't help, they didn't come. [Jingzhou Zhao] 14:22:56 You know, another appreciation one. Alright, stay on. [Jingzhou Zhao] 14:23:11 And a comment of understanding, I think. And so then we're getting talented. [Jingzhou Zhao] 14:23:23 Well, you… [Jingzhou Zhao] 14:23:30 And of course, they are on, huh? Put Future. [Jingzhou Zhao] 14:23:37 Under theta. [Jingzhou Zhao] 14:23:46 I was just wondering why that… well, large video. [Jingzhou Zhao] 14:24:08 Non-cdr, took a rugo. [Jingzhou Zhao] 14:24:21 The Xinjiangle. Potential highway call, but it's, uh… We will kind of. [Jingzhou Zhao] 14:24:41 And the opportunity. Kind of a committee, that's a mashrova. [Jingzhou Zhao] 14:24:51 You know, from deployment. Or faster figure. [Jingzhou Zhao] 14:25:00 Cost-attract figure, IOT autofill. [Jingzhou Zhao] 14:25:16 Oh, Fernandeber… Chadam Dawson. [Jingzhou Zhao] 14:25:22 The angelanik… No, please. [Jingzhou Zhao] 14:25:29 Um, Shanda. [Jingzhou Zhao] 14:25:39 Well, we did take Francie Merito? [Jingzhou Zhao] 14:25:44 Just how you're working on a dimension. Until the men drove. [Jingzhou Zhao] 14:26:05 So it's like a narcotic, uh… [Jingzhou Zhao] 14:26:11 Well, those are… they chime in. Once she says that, it's an opportunity to go. [Jingzhou Zhao] 14:26:18 Once I download that, you can send it to time for Yan. [Jingzhou Zhao] 14:26:26 Um, Japan… It's, uh… [Jingzhou Zhao] 14:26:37 Well, mahaha. [Jingzhou Zhao] 14:27:04 Well, it's… Oh, I would… We all do spotted in here. [Jingzhou Zhao] 14:27:25 Lab furniture. [Jingzhou Zhao] 14:27:43 There you go. [Jingzhou Zhao] 14:27:54 So you can tell me about your channel, which might be other… Sure, sure. [Jingzhou Zhao] 14:28:16 Your yoga lot, Isha, your yoga logician. I'm like, a paper. [Jingzhou Zhao] 14:28:22 Some nights ago, either. And from this year, um… [Jingzhou Zhao] 14:28:44 It's a kind of wish it a reminder. Sure, that should be… Invania. [Jingzhou Zhao] 14:29:07 We shouldn't hold everybody kind of way. [Jingzhou Zhao] 14:29:16 Go ahead and the individual builders. [Jingzhou Zhao] 14:29:30 But, I know, during your BR, [Jingzhou Zhao] 14:29:43 Lilia. During the TDR release month. [Jingzhou Zhao] 14:29:54 Have a look at. Hannon Tracker. [Jingzhou Zhao] 14:30:27 Ooh. [Jingzhou Zhao] 14:30:33 Whatever your sense and balance your balance over it, and the necessity as you can have a lot of them. [Jingzhou Zhao] 14:30:40 Go ahead. [Jingzhou Zhao] 14:30:51 You're prey on by now, shouldn't I? Some tuna inkasha. [Jingzhou Zhao] 14:30:58 So then… [Jingzhou Zhao] 14:31:13 But for the whole? [Jingzhou Zhao] 14:31:18 Proactive, and… [Jingzhou Zhao] 14:31:24 Panama. [Jingzhou Zhao] 14:31:36 You can agree on… boundary. [Jingzhou Zhao] 14:31:51 Stay on the higher… I mean, the R&D menu. [Jingzhou Zhao] 14:32:05 Yeah, that's right. We would try to load them at the chance to cover it. [Jingzhou Zhao] 14:32:11 Long spot switch between Shida. [Jingzhou Zhao] 14:32:17 Those are TDR. Than I do think. [Jingzhou Zhao] 14:32:24 Very fully in there. I'm sure. [Jingzhou Zhao] 14:32:30 It's a much cleaner truly rocket. [Jingzhou Zhao] 14:32:45 Somehow. Somehow. [Jingzhou Zhao] 14:33:05 The shorter side. And that is what I'll show you about. [Jingzhou Zhao] 14:33:16 Special Woman. Yeah, short answer. [Jingzhou Zhao] 14:33:25 And the challenges that we don't have on TPZ129 density of GI ruler, so we… called TPC. [Jingzhou Zhao] 14:33:36 So that must be all the functional competition in the football game under… [Jingzhou Zhao] 14:33:43 And the ITK main ga, ITK OTK, [Jingzhou Zhao] 14:33:52 To a gentleman in short shots. Some kind of complaint. [Jingzhou Zhao] 14:33:59 Employee shares, from Juba. I was talking in the end… [Jingzhou Zhao] 14:34:10 As sure to track the whoever these conditions here. [Jingzhou Zhao] 14:34:24 Would you first choose some colleagues are even in India. For the ITK, OTK? [Jingzhou Zhao] 14:34:36 Sort of do… [Jingzhou Zhao] 14:34:41 Those requests for the Chinati National. Tracking the UTV. [Jingzhou Zhao] 14:34:53 Yeah, so that's what's back in. Let me share. Commonwealth, permanent one. We also just that the whole team goes up. [Jingzhou Zhao] 14:35:01 Okay, you can count how many users? [Jingzhou Zhao] 14:35:20 So I firmware it? Do I have this one right there? [Jingzhou Zhao] 14:35:22 As a formula, Bill. [Jingzhou Zhao] 14:35:37 That's always a really money, though. When your sentences and go track it. [Jingzhou Zhao] 14:35:46 Chitan, it's easier, the whole… [Jingzhou Zhao] 14:35:55 Down Hotagayu. [Jingzhou Zhao] 14:36:02 Okay, you don't need them. [Jingzhou Zhao] 14:36:07 I can do that. It's in the Sanger Boscher. [Jingzhou Zhao] 14:36:16 In the show for the… or you vote. [Jingzhou Zhao] 14:36:27 Who made your comments? [Jingzhou Zhao] 14:36:35 To continue on. Sure, sure, sure. Sounds really personal. [Jingzhou Zhao] 14:36:47 Um, and then turned out, I will do the usual money just trying to do. [Jingzhou Zhao] 14:36:51 Table phone. [Jingzhou Zhao] 14:37:00 Thank you. [Jingzhou Zhao] 14:37:08 For your grandfather and how you were doing something good. [Jingzhou Zhao] 14:37:20 On each other. Let's still… so… publish shortly. [Jingzhou Zhao] 14:37:39 And how many yuan is… [Jingzhou Zhao] 14:37:54 I had just a demo. The ecotographer Agenda. [Jingzhou Zhao] 14:38:04 $2,000, Soyi… [Jingzhou Zhao] 14:38:15 Conscious that open TV so much someday. So, to be around… [Qidong Zhou] 14:38:33 If you run a rigor, and you measure a… Psalm 1, The GPU number has 0. [Qidong Zhou] 14:38:44 You don't been paid on numbers here. So far… so far, to be the careful. [Qidong Zhou] 14:38:52 To trigger GPU. [Jingzhou Zhao] 14:38:56 One. [Qidong Zhou] 14:39:04 There's a beautiful. [Jingzhou Zhao] 14:39:10 They all do others. Tpc? [Jingzhou Zhao] 14:39:26 Um… [Jingzhou Zhao] 14:39:40 Which is not happening yet. [Qidong Zhou] 14:39:48 When you measure… Usually your audience… [Qidong Zhou] 14:40:12 There's a result, uh… Due to a registration amount. [Jingzhou Zhao] 14:40:12 I'm eating, I'm eating. [Jingzhou Zhao] 14:40:21 Rita Twin CPU, lah. [Jingzhou Zhao] 14:40:30 You are a love animal justice for you to come down. [Qidong Zhou] 14:40:37 Actually, the CPU to point out as well. [Qidong Zhou] 14:40:48 Uh, Pokoida is your gentleman. Uh, CPC Offline the CPC that, uh, pointed out from? [Jingzhou Zhao] 14:40:59 Great. Sw. [Qidong Zhou] 14:41:09 Gpu boundary. You merely, I wanna… [Jingzhou Zhao] 14:41:17 Noah. [Jingzhou Zhao] 14:41:29 So now you… [Jingzhou Zhao] 14:41:42 This is Jiam and Li Xian, Ranjan Suki. [Qidong Zhou] 14:41:50 Oh my god. Continue to go read allergies, we don't care that now, if you're… the offline integral and the point allocated items, we all, you know. [Qidong Zhou] 14:42:07 Venice. [Qidong Zhou] 14:42:14 Dingophile, and it was strategically, that is, uh… [Qidong Zhou] 14:42:25 Yeah, sure enough for attaching it. [Jingzhou Zhao] 14:42:28 Great. [Jingzhou Zhao] 14:42:41 Give me a… [Qidong Zhou] 14:42:57 To share the… [Jingzhou Zhao] 14:43:08 Italk. [Qidong Zhou] 14:43:13 To talk. [Qidong Zhou] 14:43:33 Patrick, uh… IMEO GPU. [Qidong Zhou] 14:43:39 Gpu, [Jingzhou Zhao] 14:43:49 So, ma'am. I am. [Qidong Zhou] 14:43:54 Xin Pei. [Jingzhou Zhao] 14:44:02 But in particular, I'm CPCSW. [Jingzhou Zhao] 14:44:08 Now I'm gonna read that, huh? [Qidong Zhou] 14:44:17 Uh, yeah. [Qidong Zhou] 14:44:31 The NYU is here to… Yeah, engine generic, uh, somewhat, yeah. [Qidong Zhou] 14:44:40 You can get up, actually. It's okay, Simon. [Jingzhou Zhao] 14:45:09 Then you can see the zone. And have… put 100 inches as well. [Jingzhou Zhao] 14:45:24 Thomas? 2010? [Jingzhou Zhao] 14:45:49 May happen. Oh, one day right after… oops. [Jingzhou Zhao] 14:45:56 Junior showed you, don't. [Jingzhou Zhao] 14:46:13 What's it wrong? [Jingzhou Zhao] 14:46:50 Hey, can you visit a journal? [Jingzhou Zhao] 14:46:55 Oh, that's true. [Jingzhou Zhao] 14:47:00 Let's see… [Jingzhou Zhao] 14:47:17 Hang on, yeah. Well, Ms. Professionalism. [Jingzhou Zhao] 14:47:26 Let me pull this up. [Jingzhou Zhao] 14:47:32 You don't want. [Jingzhou Zhao] 14:47:44 Manager with each other. We understand that track in the joggle. [Jingzhou Zhao] 14:48:08 What's come down? Can't bog up? [Jingzhou Zhao] 14:48:25 How the How many OEs here? [Jingzhou Zhao] 14:48:40 How can you not do that? [Jingzhou Zhao] 14:48:56 Balanced yet? Why don't you admit that? [Jingzhou Zhao] 14:49:10 Me on. [Hongyu Zhang] 14:49:16 目前没有这个，我们的会跟他们是次次重，我们只能去看他的那个。 [Hongyu Zhang] 14:49:22 每次梁志军给出来的那个ppt。 [Hongyu Zhang] 14:49:26 但是这个信息好像更新的，也没有什么特别新的东西。 [Hongyu Zhang] 14:49:34 对他们一直是周四下午跟我们完全是宠的。 [Jingzhou Zhao] 14:49:37 I didn't have time. [Hongyu Zhang] 14:49:46 看看他们有没有什么变化呢？ [Hongyu Zhang] 14:49:49 另外一个就是我们就主动的去跟人家去了解。 [Hongyu Zhang] 14:49:53 老这么被动的听，大概现在也也发现是不行。 [Jingzhou Zhao] 14:49:54 Great. [Hongyu Zhang] 14:50:03 对，对，对，对，对，对。 [Jingzhou Zhao] 14:50:11 Uhdr. Well, I'm gonna say you have any comment. [Jingzhou Zhao] 14:50:18 I want to show you value some of this. And the U, Yarma. [Jingzhou Zhao] 14:50:26 I'm a 9 out of 10, how many count for? [Hongyu Zhang] 14:50:30 对我们也是。找个时间可以约他们去真的了解了解，先看完了他们的。 [Hongyu Zhang] 14:50:36 一系列报告吧，那个他们报告确实是梁志军的风格，就是每次都把一个同样的报告放在那儿。 [Hongyu Zhang] 14:50:53 对对。 [Jingzhou Zhao] 14:50:54 It's in Hamlet. The whole… the Indians in Hamilah. [Jingzhou Zhao] 14:51:00 Tbc. 200, ma'am. [Hongyu Zhang] 14:51:05 对，我觉得你刚才那个讲的有有点侧重，分工可能。 [Hongyu Zhang] 14:51:10 就是大家去了解的就是更更针对性一点儿。比如说。 [Hongyu Zhang] 14:51:15 像这种顶点或者是akotk的，如果是徐唱名。 [Hongyu Zhang] 14:51:20 确定的话，那就可以多花点儿功夫去做这个事情。 [Jingzhou Zhao] 14:51:27 3. Go to the top. [Hongyu Zhang] 14:51:30 对光协议的话，我觉得这个毕不了业了，都。 [Jingzhou Zhao] 14:51:35 We also know… Asana. [Hongyu Zhang] 14:51:42 对，现在就是，虽然你重复他做的东西非常辛苦。 [Hongyu Zhang] 14:51:47 但是他也不够我的感觉。 [Hongyu Zhang] 14:51:56 对，至少是针对我们这个在做这个就不是完全的follow别人的东西。 [Jingzhou Zhao] 14:52:02 But… but… [Jingzhou Zhao] 14:52:06 To him today. A little harder. [Hongyu Zhang] 14:52:06 对，然后对，对，对。 [Jingzhou Zhao] 14:52:21 I mean, I need to go in the team dynamo? [Jingzhou Zhao] 14:52:28 Huh? [Jingzhou Zhao] 14:52:37 We'll play… [Jingzhou Zhao] 14:52:48 It's a government here. Good receivers. [Jingzhou Zhao] 14:52:55 So you can pull off now. It's a message. [Jingzhou Zhao] 14:52:59 Show you as a message, too. Someone to quanner. [Jingzhou Zhao] 14:53:06 So it's a new negative abandonment bar. Sure. [Jingzhou Zhao] 14:53:16 I mean, as a… At last day, you're getting a homeless, for the new one, you can start English. [Jingzhou Zhao] 14:53:30 In the level of car. [Jingzhou Zhao] 14:53:49 I think they just assumed. [Jingzhou Zhao] 14:54:57 So, gentlemen is watching me. [Jingzhou Zhao] 14:55:04 I should say, hey, you can kind of edit the LCP, it's kind of like a GPU funds.com. [Jingzhou Zhao] 14:55:09 Rcp, the contact. If-else. [Jingzhou Zhao] 14:55:21 To Zoom units feeding. That's a… [Jingzhou Zhao] 14:55:33 Cool. Yes, that was the flat design. [Jingzhou Zhao] 14:55:57 The Malay area. [Jingzhou Zhao] 14:56:03 Well, conversation. I mean, not there. [Jingzhou Zhao] 14:56:13 Forming the floor. [Jingzhou Zhao] 14:56:25 Sure, no pressure. So now… Oh, not even shot a little bit of that. [Jingzhou Zhao] 14:56:35 Not really, no, no, no. So what's that? [Jingzhou Zhao] 14:56:41 Sure, you're not him at all. Cool. I'm not careful. [Jingzhou Zhao] 14:56:47 Through audio. In other words, so… Sure. [Jingzhou Zhao] 14:57:00 So we created a shock. No. Oh. [Jingzhou Zhao] 14:57:07 Shaduruma. [Jingzhou Zhao] 14:57:22 Aren't you the… Anima… [Jingzhou Zhao] 14:57:40 But I will not update. I don't have Angela full. [Jingzhou Zhao] 14:57:58 Being a surge of surgeon. The RD, if I know. [Jingzhou Zhao] 14:58:05 Father will pass. [Jingzhou Zhao] 14:58:10 The… I'm not being… [Jingzhou Zhao] 14:58:37 They opened it up. Down the specific… We don't have numbers. [Jingzhou Zhao] 14:58:51 Yeah. So we've had a… Huh? [Qidong Zhou] 14:58:54 Okay? [Jingzhou Zhao] 14:59:03 So, um… [Qidong Zhou] 14:59:14 Well, bro, you send angelieber. How's it going? How's your… [Qidong Zhou] 14:59:27 Maintain hard for your chin. [Hongyu Zhang] 14:59:30 那上次秦晖荣专门在那儿说是一年3000欧，她觉得这数儿还挺多的。 [Jingzhou Zhao] 14:59:43 To see you. There is some. [Hongyu Zhang] 14:59:45 不不知道是不是已经要交了还是怎么着，反正是。 [Qidong Zhou] 14:59:51 Devin? [Jingzhou Zhao] 14:59:57 As CB… [Qidong Zhou] 15:00:10 Yeah, much energy. [Jingzhou Zhao] 15:00:18 Well, the… [Jingzhou Zhao] 15:00:31 3 years will attach. [Jingzhou Zhao] 15:00:49 Okay, so this time… [Jingzhou Zhao] 15:00:58 It's all. Beautiful. Okay, okay. [Jingzhou Zhao] 15:01:04 So let's, uh, start English art. Um, before in the Chinese part, we just, uh, introduced our current status leaders. [Jingzhou Zhao] 15:01:15 Of the TR, uh, document. Uh, our document, uh, almost, uh, finished. [Jingzhou Zhao] 15:01:22 And, uh, we also… optimize, uh, features. [Jingzhou Zhao] 15:01:30 From, uh, Jetpack or Kinji to PDF, Uh, with, uh, high, uh, high-quality pictures for publications, so… Until now, uh, only, maybe only one picture needs to update. [Jingzhou Zhao] 15:01:49 That may be able to finish our… works for the TDR, but we still need to wait Uh, the… feedback from the inside review, um. [Jingzhou Zhao] 15:02:05 To get the feedback, then we must, uh, then we can… modifying our… uh, worse. Total the… bit of back. [Jingzhou Zhao] 15:02:19 Anyway, the TDRs work. Almost to finish, and we will, uh. [Jingzhou Zhao] 15:02:28 Moved to, um… PDR forward TDRs, uh, guideline, okay. [Jingzhou Zhao] 15:02:37 Let's see how much is it? [Jingzhou Zhao] 15:02:48 8… We will see my slide? [Jingzhou Zhao] 15:02:56 Okay, so in our TDR, that's the final summary part. Uh, I just wrote 3 directions. [Jingzhou Zhao] 15:03:09 For R&D, first is simulation and trigger algorithm. Uh, it's need many people to join, and the quality later to find some, uh, useful algorithm and solution. [Jingzhou Zhao] 15:03:24 Uh, for CPC, uh, reference TDR. And second is for the… Uh, trigger hardware parts. [Jingzhou Zhao] 15:03:32 We will, uh, developed a prototype, uh, common board. And, uh, build a prototype test back. [Jingzhou Zhao] 15:03:42 Well, our, uh… part of communication, and uh… Uh, trigger structure, and also, uh, we can test, uh, uh, the algorithm. [Jingzhou Zhao] 15:03:54 That's this hardware part, uh, hardware… uh… platform. [Jingzhou Zhao] 15:04:01 And a third one is we… we should study, uh, the distributed parent uh, computing technology. [Jingzhou Zhao] 15:04:10 It's just a general description, but we want to… Um, terabyte per second level, uh, DAC system. [Jingzhou Zhao] 15:04:22 Because as the, uh… Before CPC, uh, we… Just, uh, working on, uh… 10 gigabytes per second level, so we… we try to increase the data uh, rates, uh, requirement for CPC. [Jingzhou Zhao] 15:04:41 And also, we want to… Uh, try RDMing protocol, and also by the GPU-based acceleration. [Jingzhou Zhao] 15:04:51 For online, uh, online software, and also for the high-level trigger. [Jingzhou Zhao] 15:04:57 So this third, uh, study directions. But this is only for the TDR. [Jingzhou Zhao] 15:05:05 And we also want to… talk more about our detailed plan for the next, uh. [Jingzhou Zhao] 15:05:14 Tool or three, uh, two, two or three years. Uh, this… we… visual discussion. [Jingzhou Zhao] 15:05:21 So, this is a general introduction for our uh, discussion, uh, before… Uh, before Chinese parts. [Jingzhou Zhao] 15:05:32 And now we can go through to the… different part, uh, progress. [Jingzhou Zhao] 15:05:40 But for the RDMA, just… just a moment. There are two, uh, firmware release, but still on the debugging. [Jingzhou Zhao] 15:05:52 So, there are no, uh… progress to report, but I think maybe next one or two. [Jingzhou Zhao] 15:06:01 Meeting, we can got some… Fumer Report. [Jingzhou Zhao] 15:06:07 From the RDMA. And for the trigger ball. [Jingzhou Zhao] 15:06:15 Window and tongue. Still, uh, working further. [Jingzhou Zhao] 15:06:20 Uh, schematic… plant for the common board. [Jingzhou Zhao] 15:06:25 Uh, maybe they still need some time to do that. There are no much progress, that's… very recent… recent time. [Jingzhou Zhao] 15:06:37 So, the first is… [Jingzhou Zhao] 15:06:45 Okay? And, uh, she, um, have another sense for the tune also. [Jingzhou Zhao] 15:06:53 For the TPC part, we will talk Okay, next time. [Jingzhou Zhao] 15:06:58 So we move to the simulation. Progress from booking. [Jingzhou Zhao] 15:07:04 I have a slide for the… chat. Okay, you can share your screen. [Jingzhou Zhao] 15:07:22 Or, as a… [Jingzhou Zhao] 15:07:29 Some, uh… But this COVID has tried, uh, some… track algorithms. [Jingzhou Zhao] 15:07:37 Average intent, there are results in the… So this is a very brief introduction about the check. [Jingzhou Zhao] 15:07:45 Is, uh, Alex. Alex, uh, for the charged particle inside the… So, the SY Pen is a circle. [Jingzhou Zhao] 15:07:58 In the S3 or YZ10 is a… super non-metric function. [Jingzhou Zhao] 15:08:08 Rsim or cosine functions. What's on the top rod. [Jingzhou Zhao] 15:08:13 And the functions for this car is, uh… these three pictures, sorry, what's the dimensions? [Jingzhou Zhao] 15:08:24 All the diameter is only for IPK, or… So, it's a headache. [Jingzhou Zhao] 15:08:36 Rocha and Shen. You mean Magnetek. Magnet. Yes. [Jingzhou Zhao] 15:08:41 So inside the magnitude. All the part… all the charged particles inside, uh. [Jingzhou Zhao] 15:08:46 Magnetic… magnetic field. The trajectory is Alex. [Jingzhou Zhao] 15:08:51 And this is the two equations that describe these, uh, highlights. [Jingzhou Zhao] 15:08:56 So there's assumptions that the… to start from their IP. [Jingzhou Zhao] 15:09:00 So, yeah, this background, back is a poacher at 000. And the case, the curvature, 1 over the radius. [Jingzhou Zhao] 15:09:14 Pi-zero is the initial angle. For the track, and… That's the head example. [Jingzhou Zhao] 15:09:23 Access to occupants. Uh, the SY10. [Jingzhou Zhao] 15:09:30 So… Um. [Jingzhou Zhao] 15:09:34 We say the trade reconstructions, you can use these functions to fit the check. [Jingzhou Zhao] 15:09:39 For the hardware, it's difficult to do their… teaching. So… so… As far as I know, there are two main… Messer for the tracked sugar in the hardware. [Jingzhou Zhao] 15:09:54 Triggered one is better call the off-transform, the other one is the pattern recognition. [Jingzhou Zhao] 15:09:58 So, I tried to off-transform these fields, so it's… Uh, it transformed… transformed the 3D… it transformed the information from the 3D spatial coordinate. So in our case, it's the… Simi Drisco coordinate into a parameter space. [Jingzhou Zhao] 15:10:24 Then, in the parameter space, each key, each point, the letter represents a… 3D curve. [Jingzhou Zhao] 15:10:33 The premise, but it's… so, for example, on the right side, it's the… It's a… in the track. [Jingzhou Zhao] 15:10:44 Detect her, and each tick will be transferred to the parent. [Jingzhou Zhao] 15:10:50 So, like, like, the prop… here. So, this is the… this proper sender tracking in the 3D spatial coordinator. Coordinate. [Jingzhou Zhao] 15:11:01 And each point here is represent a heat. And the path transform will transfer their information for this heat into the parameter space, and in the parameter space, he comes a curve. [Jingzhou Zhao] 15:11:15 That this one here represents, uh. Occur in the right side. [Jingzhou Zhao] 15:11:21 And so, if there are 6 feet in the 3Ds. Spatial coordinates, then we will get 6. [Jingzhou Zhao] 15:11:30 Curve, and uh… This project sends the check Anyways, and uh… parameters for this track candidate can be is the, uh… The SY riders, uh… has zero render. [Jingzhou Zhao] 15:11:48 2 divided by PT at this point. So, with this half transform, you can… find the check ID, and also get the… parameters for this check-in day. [Jingzhou Zhao] 15:12:01 And… Um, usually there would be assumptions when we apply the off-transform. [Jingzhou Zhao] 15:12:08 This is the, uh… not PTAPT assumptions. [Jingzhou Zhao] 15:12:14 Assume that the PT is larger than some certain value, for example, 23V. [Jingzhou Zhao] 15:12:19 These assumptions, the angle. His angle would be small. [Jingzhou Zhao] 15:12:24 So the sine prime minus prime 0. It's a… uh… It's almost April is because to the 5… 5 times 5-0. [Jingzhou Zhao] 15:12:37 And the Z, uh, becomes the radius times the tangent number. [Jingzhou Zhao] 15:12:44 So, these functions [Jingzhou Zhao] 15:12:50 Can be used for their, uh, off. Can be used for the, uh, track constructions. [Jingzhou Zhao] 15:12:57 Um. So, as for the detector, so I try to… use both the informations for the Word test, you know, so ITK. [Jingzhou Zhao] 15:13:08 But for the warehouse, for the innermost two-layer, there are too many information, too many… background teeth. [Jingzhou Zhao] 15:13:17 So, I only used the four areas. This tool and, uh, this gift. [Jingzhou Zhao] 15:13:24 So, for these four layers, the background is much… That's a terrible story. [Jingzhou Zhao] 15:13:32 And also, the information is three layers. Well, possibly 7, 7 layers information. [Jingzhou Zhao] 15:13:40 Um, so that's one thing. This picture, it's still… embedded 10 punchcrossing. [Jingzhou Zhao] 15:13:48 Background together. I try post one bunch and the 10 bunch. [Jingzhou Zhao] 15:13:52 Okay, let's summarize all the results here. So… Yeah, so first I use only an RCM on frame, and try… So it involves also the workouts, the outermost four-layer for work tasks. [Jingzhou Zhao] 15:14:06 Elzar tried their… 10 bunch, and one bunch for you, for the background. [Jingzhou Zhao] 15:14:12 So, yes, there's a summary. So, for example, for the ITK and OTK only. [Jingzhou Zhao] 15:14:19 We have 10 bunch crossings. For one background event. [Jingzhou Zhao] 15:14:23 Background, if you're showing me. 100% sense. I mean, it's all the backgrounds are asked and selections. [Jingzhou Zhao] 15:14:30 And if you are using the one-point passings. Uh, it's only in Paragua. [Jingzhou Zhao] 15:14:36 The fibrom efficiency is 40%, still very, very large. Uh, basically include also their OTAs. [Jingzhou Zhao] 15:14:45 Yes, only one bunch. Crossing the background intrusion system. [Jingzhou Zhao] 15:14:49 1%. This is acceptable. [Jingzhou Zhao] 15:14:56 Okay, so use… So your conclusion is Vortex is… We will contribute very good efficiency in IT and OK. [Jingzhou Zhao] 15:15:09 With also the work has… or what has crossed ITK and… I mean, if only use the ITK broadcaster, if… Efficiency is bad. [Jingzhou Zhao] 15:15:21 Let's add additional vertex is… it's better. Yes. Only ITK and UK, there are only four years. [Jingzhou Zhao] 15:15:32 I think for measuring, it's just not enough. Google providing a check. So with another four-year property. [Jingzhou Zhao] 15:15:42 Part 7, sorry, this trader. Totally eight layer. 8. [Jingzhou Zhao] 15:15:46 Player than… then we can gather… equip your constructions for the check. [Jingzhou Zhao] 15:15:56 But the edge efficiency is not true. It's only 93.5. [Jingzhou Zhao] 15:16:10 You should compel with, uh… Both algorithm is e-call, edge call. [Jingzhou Zhao] 15:16:18 In college, crap. Give me the caramaker. Yeah, yeah. [Jingzhou Zhao] 15:16:23 I mean, this efficiency, we must compare to the Uh, you can't random. [Jingzhou Zhao] 15:16:30 That's cost efficiency. We seem… a single event. [Wolfgang Kuehn] 15:16:39 Sorry, can I ask a question? I mean, definitely you will not run an experiment where you only want to take data which have two neutrinos and two muons from the Zhix decay. So, I mean, there will be many other parallel triggers, so this less than 100% deficiency is no problem, I think. [Jingzhou Zhao] 15:16:41 Yes? [Jingzhou Zhao] 15:17:04 Sorry, to be… [Wolfgang Kuehn] 15:17:04 I mean, you can never, with such a trigger, you can never achieve 100% efficiency. [Jingzhou Zhao] 15:17:13 Your current records. Slightly longer. [Wolfgang Kuehn] 15:17:16 Yeah, but you… I mean, you will not run the experiment in the way that you only take these events. There will be many other events which will be considered, other physics events, and so even if the efficiency for this trigger is not 100%, there may be another physics trigger which gets this event through, huh? So, I think… This is to… this kind of view is too one-dimensional. I think you should look [Jingzhou Zhao] 15:17:17 What do you say? [Wolfgang Kuehn] 15:17:40 At a mix of triggers for various physics aspects, and then the simulation will be very different than if you just simulate this single trigger. Of course, I mean, if you just have 8 layers of silicon or so, and you require two muons, the efficiency can never be 100%. [Wolfgang Kuehn] 15:17:58 Because the detectors have an intrinsic inefficiency, and also the algorithm doesn't work 100%, so I think this is asking for too much to have 100% or 99.5 percent efficiency. [Wolfgang Kuehn] 15:18:13 I think you have 39… 39% is very… is a very good number, I think. [Jingzhou Zhao] 15:18:13 And we've… [Jingzhou Zhao] 15:18:20 Actually, with the carameter and the neuron detector, these two sub-detectant information remotely gather 99% for I can add a point. [Wolfgang Kuehn] 15:18:27 Yeah, sure. If you combine different detectors, then the efficiency will get larger, yeah, right? So it's not… there is no point in just looking at a single trigger and its efficiency and say, oh, it's not good. [Jingzhou Zhao] 15:18:30 That's what happened. [Wolfgang Kuehn] 15:18:38 Because you have to look at the complete trigger information. And so there are complementary information, so if you have a hit in the muon detector and you don't have the hit in the In the tracker, because your algorithm fails for this track, then you still get the event. So I think [Wolfgang Kuehn] 15:18:54 One should not look at isolated trigger signals, and then complaining that the trigger efficiency is not high enough. [Wolfgang Kuehn] 15:19:02 You have to look at the combination of all trigger informations. [Jingzhou Zhao] 15:19:06 Yes, yeah, we've got a… I can make a… type of global sugar efficiency. [Wolfgang Kuehn] 15:19:15 Right, that is the relevant number. What is the total efficiency that the combination of your triggers with the whatever tracking, colorimeter, immune detector, uh, triggers that event. This is the only relevant number, I think. [Jingzhou Zhao] 15:19:37 Yeah, but uh… the current… So if… for the current average, I'm sorry. [Jingzhou Zhao] 15:19:45 Using only the kilometers and the mirror detector efficiency for the physical processes, or the 99% So, I think the pack trigger We don't see any improvement. [Wolfgang Kuehn] 15:19:51 Yeah, right. [Wolfgang Kuehn] 15:19:56 No, no, no, sure. Yeah, I mean, but that is actually what you want to know. What is the efficiency that a certain physics event is triggered, considering all the trigger information which you have? [Wolfgang Kuehn] 15:20:05 I think this is the relevant one number, and there is no point in saying, oh, 39%… 93% for the track trigger is not good enough. This is not the… this is not really the point, I think. [Jingzhou Zhao] 15:20:23 Ees, uh, I also think you must come and not come tonight. [Jingzhou Zhao] 15:20:32 This, uh… Former agreedym together. [Jingzhou Zhao] 15:20:36 To know how the… Uh, before the background rate could be reduced to 3.5%. [Jingzhou Zhao] 15:20:45 Right? Which means about close to the 50 kilohertz. Yes. If we add this track trigger. [Jingzhou Zhao] 15:20:53 Agrees all together, so the… background reach could be reduced to… Do we increase the air bunker? [Jingzhou Zhao] 15:21:04 Increase. That's this. All. Because either this one or that one. [Jingzhou Zhao] 15:21:15 So, in future, how to deal with the different trigger primitives… Uh… in different, uh, detectors. [Jingzhou Zhao] 15:21:30 All the logic, or… and the logic. [Jingzhou Zhao] 15:21:39 So, Karen? Yeah. Account for all. [Jingzhou Zhao] 15:21:45 Oh, yeah. We're down to N. [Jingzhou Zhao] 15:21:51 But for some process, like his tokamagama, there's no track. It won't pass this kind of a trigger. [Jingzhou Zhao] 15:22:01 Yes, so one after your… Oh, so one physical event, you will, uh… decided by only one. [Jingzhou Zhao] 15:22:11 Detector. Sometimes we should make a different, uh. [Jingzhou Zhao] 15:22:20 Conditions set. Why is only detector's condition? One is combined nation's condition that's different. [Jingzhou Zhao] 15:22:28 Okay, okay. If the entity is too big. [Jingzhou Zhao] 15:22:34 Then we can, you know, use the… I don't mean to her directly. [Jingzhou Zhao] 15:22:41 Uh, so for that… for this? True, the main part is to reduce the background, or… Select more signal. [Jingzhou Zhao] 15:22:51 If we want to reduce it back on time, we use N. If we want to set up more pack for single, we use all C. [Jingzhou Zhao] 15:22:59 I see. Because originally, I thought the chat sugar used to… it's a lot more… that, uh, you know, Facebook, or some low-energy events, that's why we had. [Jingzhou Zhao] 15:23:13 Had set up, but… I can try and… You can see how much it crackers. [Miroslav Saur] 15:23:20 Sorry to jump in, but, well, in general, what is the purpose of Tega? [Miroslav Saur] 15:23:26 To both reduce background and select the signal as efficiently as possible. [Miroslav Saur] 15:23:31 So, as I mentioned before, mo… the… the most efficient way will be some kind of combination, maybe based on Additionally information, yeah? So maybe if… information in kilometer is above some level, like… emerging… deposit emerging kilometers above some level. [Miroslav Saur] 15:23:51 Then it can be kilometer on its own. If energy is below some level, then maybe it would have to be kilometer and something. [Miroslav Saur] 15:24:01 Yeah, so… Uh, so possibly to really be efficient and… be efficient on signal, and at the same time to surpass back-end as much as possible. [Miroslav Saur] 15:24:11 The simple logic is, uh… is not enough, yeah? So I… I would say one alias has to go into more complex logic, and then, of course, is the… is the question, how much of logic gonna be, uh, put on a hardware level to be still fast enough to make all the evaluations, yeah. [Miroslav Saur] 15:24:33 In the limits here for hardware, they go. [Jingzhou Zhao] 15:24:45 I… [Wolfgang Kuehn] 15:24:45 I think you… what you have to do is you have to make for… for each physics signal, you have to make a plot Efficiency versus purity. [Wolfgang Kuehn] 15:24:56 Of the trigger selection. And then you can play with your various, as was just mentioned, can you play These are various ingredients into the trigger decision, like Is there other tracks? How many tracks? Is there a colorimeter signal? How high is it? [Wolfgang Kuehn] 15:25:10 Is there a muon detector, and so on. This… and then you… you just vary your parameters, and then you can see how well you are doing in this matrix. [Wolfgang Kuehn] 15:25:20 Efficiency versus purity of the signal, and then you can optimize the trigger. [Jingzhou Zhao] 15:25:29 Right. [Jingzhou Zhao] 15:25:33 Okay, so this, uh… preliminary results, just only use, uh. [Jingzhou Zhao] 15:25:41 Track detector, so… [Wolfgang Kuehn] 15:25:47 Yeah, it's one input, it's one… as was discussed before, I mean, it's an input to the final trigger decision. [Jingzhou Zhao] 15:25:48 As… [Wolfgang Kuehn] 15:25:55 But it probably will never be the only trigger a decision for any physics event. You want to look at other detectors, because you… have fast detector information available, so you want to use everything which is fast enough, so everything except for the TPC, [Wolfgang Kuehn] 15:26:10 You can essentially use, and you should use it, yeah? [Jingzhou Zhao] 15:26:20 You use, uh… single sample statement is, uh… all the efficiency. [Jingzhou Zhao] 15:26:28 Yeah, same… same… Okay, now you can… can mesh… Between the… all the algorithms. [Jingzhou Zhao] 15:26:37 Messed up. [Qidong Zhou] 15:26:46 Bobin, can you tell me, uh, how you define the, um, you have using the vertex, for example, ITK or OTK? [Qidong Zhou] 15:26:55 How you define that, um… For example, on the pixel would detect a signal. [Qidong Zhou] 15:27:03 It's just, um, did you consider… I mean, for the simulation point of view. [Qidong Zhou] 15:27:10 Uh, is the simulation currently now for ITT and OTK has the digitalization information in each pixel. [Qidong Zhou] 15:27:20 I mean, that's how you define the heat on the ATK or OTK. [Qidong Zhou] 15:27:27 So, I'm wondering how you define this heat. [Jingzhou Zhao] 15:27:30 I would say financial resource from the… from both sides of the pay anymore. [Jingzhou Zhao] 15:27:37 But there are four layers. [Qidong Zhou] 15:27:38 I mean, look… One pixel was fired, or you needed several pixel has some, uh… Um, he… he to count. [Jingzhou Zhao] 15:27:49 Yeah, that's, um… So far, just for the issue. [Jingzhou Zhao] 15:27:55 Pixels. So, as long as it's fire, you'll see… [Qidong Zhou] 15:28:02 So just the one pixel file that is okay, or you need some combination kind of cluster. [Qidong Zhou] 15:28:09 Definition. Did you… [Jingzhou Zhao] 15:28:10 That's something I want to do next. So, so far, the input is the heat. [Jingzhou Zhao] 15:28:17 Uh… But once I've also found that one check will generate many hits. [Jingzhou Zhao] 15:28:23 At the same locations. So, that's… I think that's where we do a faster commercial audiences here. [Jingzhou Zhao] 15:28:32 And probably other things, Karl. The duplicated removal. [Jingzhou Zhao] 15:28:37 So, you know, how transformed, it will generate are really not just one… one… maybe not just one hand, because… from just one theme past the selections, there were other teams also pass the selections. [Qidong Zhou] 15:28:55 Yeah, I think it's… [Jingzhou Zhao] 15:28:58 And this would be what I'm going to do next. [Qidong Zhou] 15:29:01 Yeah, I think it's, uh, important that you define There's some kind of cluster pattern that, uh. [Qidong Zhou] 15:29:09 Can be used as a primitive way for the L1 trigger implementation, then you… do this kind of calculation based on that single heat, uh. [Qidong Zhou] 15:29:20 Unit, as defined in the level 1 trigger. And then check this again. [Qidong Zhou] 15:29:27 Another point, can you go to the Hof transformation page? And, um, yeah, this… this is the typical Holt transformation as a 2D transformation, yet how Putin with 3D, um… Actually, this is your, um, basically is on the 2D, um, space. [Qidong Zhou] 15:29:53 At your… I suggest that you also to check in the future. [Qidong Zhou] 15:29:57 To also be in this, um… to the, uh, hope transformation. [Qidong Zhou] 15:30:03 And add another dimension with theta. Yeah. So, as this tracker inside the magnetic field, it will be helix. [Jingzhou Zhao] 15:30:05 I already actor, uh… The security is the hope transform. [Qidong Zhou] 15:30:18 So, um, on the ZD direction, you can also bend this 2D transformation with Theta and omega, and omega can be also here, the Z. [Qidong Zhou] 15:30:29 And you… you can, uh… do this kind of calculation Also, being the beta setup. [Qidong Zhou] 15:30:36 As a parallel angle of another dimension. [Qidong Zhou] 15:30:43 I'm not one… I'm not sure you will get my point. [Qidong Zhou] 15:30:47 So you have seen this here, right? You have a track with X and Y on this, uh, right-hand, uh, um, right-hand side. [Qidong Zhou] 15:30:59 You have a sixth point, and there will be 6 lines when you do a half transformation, and they cross on the point. [Qidong Zhou] 15:31:06 And, um, it's on the XY dimension, then you can also If there's, for example, signal and background. [Qidong Zhou] 15:31:17 You can also bend with another dimension, which is theta. Then you will have some, uh… doesn't match with some VIN, I mean, theta at a very small value, or That's the large value, this tracker will be on the whole transformation, uh. [Qidong Zhou] 15:31:37 Space it will be, uh, the course pointer will be across or not cross, so you can see this at the different bends. [Qidong Zhou] 15:31:45 With another pin, with a per angle for theta. [Jingzhou Zhao] 15:31:54 I'm doing 3D health transform now, by 2D… okay. [Qidong Zhou] 15:31:56 Which… [Qidong Zhou] 15:32:01 Uh, this is 3D. We also have another dimension. [Jingzhou Zhao] 15:32:04 These employees are 20-day only, because… It's difficult to control a 3D product yet, so I only target 2 viewpoints. [Jingzhou Zhao] 15:32:12 But for the… for the half-transformer, I'm doing a cerelict transform. [Qidong Zhou] 15:32:19 Oh, okay, but, uh, how you get this? [Jingzhou Zhao] 15:32:22 So, this is, uh, so the… coordinate the spatial coordinator is the city, drip codes coordinate. [Jingzhou Zhao] 15:32:32 It's R, phi, and Z. I transfer the R, phi and Z to the… the parameter space kappa by 0N number. [Qidong Zhou] 15:32:42 And this will be possible to get from, um… I think you need, uh, one, uh. [Qidong Zhou] 15:32:50 So, I think here you need a 2D tracker reconstruction. On the level 1 triggers level. So you need the algorithm to do a fitting of the tracks. [Qidong Zhou] 15:33:04 That's a… that's a… here you can get, uh, on the offline reconstruction, you can… directly get this information of, uh, maybe, uh, are you needed to… to do a filter of the track to get to the R, right? [Jingzhou Zhao] 15:33:04 Price, third. [Qidong Zhou] 15:33:22 The curve of this, uh… this, um, tracker, you need to do a fitting on the level 1 trigger algorithm to do a 2D track. [Jingzhou Zhao] 15:33:39 I didn't do the second year. This arm, no ice, the distance between the heat and the origin. [Qidong Zhou] 15:33:43 Yeah. [Qidong Zhou] 15:33:51 And that's… I'm not too sure this definition is okay. You have an OTK and a, uh… ITK, then, the R is definition of what ITK and OTK, maybe, um, that… You needed to consider about the magnetic field, and then [Qidong Zhou] 15:34:11 Do a fading of this… I mean, this two-pointer can… can also need, uh, um… a magnetic field, then you need to calculate this trajectory and to… to calculate the circle, and then you get the value of R. [Qidong Zhou] 15:34:30 So we need a 2D tracker reconstruction at the level of, uh. [Qidong Zhou] 15:34:37 You don't have this R directive. [Wolfgang Kuehn] 15:34:40 I think the Huff transformation determines the radius of the track. [Wolfgang Kuehn] 15:34:45 And this gives you the momentum. I mean, what you're saying is, uh. [Wolfgang Kuehn] 15:34:49 If you want to have a… if you do a fit, of course, you get a much better precision for the radius of the… off the track, but just to get the parameter with the course definition is just a half transformation gives you this information. I think this is, uh… [Wolfgang Kuehn] 15:35:09 There's nothing… you don't need a fit to determine the momentum of the particle. Without the fit, of course. [Wolfgang Kuehn] 15:35:15 The resolution will be worse. [Qidong Zhou] 15:35:18 Yes. So, actually, um… Um, that's… [Jingzhou Zhao] 15:35:22 Probably chose the parameter that I want to decide from this off-transform. [Qidong Zhou] 15:35:28 Yeah, this is offline. On the offline reconstruction level, you can have this But on the Level 1 figure, um… level… at the far end as I know, at the Bell 2, you need a 2D, uh, track reconstruction from the… [Qidong Zhou] 15:35:46 Track segment, and uh… And then do the transformation. [Qidong Zhou] 15:35:54 Which is necessary at the input for this off-transformation. [Jingzhou Zhao] 15:36:01 Because this 2D… this camera. This conference from… I found this in the CMS In case, uh… our business, PhD thesis. [Jingzhou Zhao] 15:36:14 And his storage is all transformed. [Qidong Zhou] 15:36:19 Yeah, I… [Jingzhou Zhao] 15:36:21 So you have quite the radius of the poetry. [Qidong Zhou] 15:36:26 Yeah, but, uh… So, on the right-hand side, there is phi and tangent lambda and R, right? [Qidong Zhou] 15:36:34 That's say you need to get. Uh, from, uh… [Jingzhou Zhao] 15:36:40 Handa is the parameter. Not the proper time, bye. [Jingzhou Zhao] 15:36:44 Zero on these three parameters. 5-0 at number. [Jingzhou Zhao] 15:36:51 An art, and he… Great. [Qidong Zhou] 15:36:51 Yeah. Can you get… can you get it from… from… I mean, from the readout of this ITK and OTK directly. [Qidong Zhou] 15:37:01 That's needed to be figured out. As an input. [Jingzhou Zhao] 15:37:06 These three companies, we need to… And we need to pay it from the power transform method. [Jingzhou Zhao] 15:37:13 The input is, uh, R, 5, and… [Jingzhou Zhao] 15:37:21 For each quant, we can get the R, bar, and V. [Jingzhou Zhao] 15:37:24 And with… from this info, we calculate… Um, we try to get the value for kappa, phi0, and lambda from the Hawk transform. [Wolfgang Kuehn] 15:37:38 I mean, the detector gives you points in space, and 3D points in space, and what the Huff transformation does, it transforms this helical shape point, depending on the parameters of the track. And so, this point tells you then what are the track parameters. So it's… [Wolfgang Kuehn] 15:37:59 It's a direct determination of the momentum of the track if you know the magnetic field. [Wolfgang Kuehn] 15:38:04 And you know the geometry of your detector. I'll buy it with a poor resolution, depending on The granularity of your half space. [Qidong Zhou] 15:38:15 Yeah, yeah, I… that's the equation that you have the moment you have the magnetic field, and the… Yeah. The resolution… Um, I'm not sure… as far as I know, at the currently running the experiment that I've built, who we need to do the reconstruction of Trek. [Qidong Zhou] 15:38:40 But I'm not sure what this case. Cms or other experiment, if you need a pre… building of a tract to get this input parameters. [Qidong Zhou] 15:38:54 I'm not sure the resolution directly from the 3D coordinator information is, uh, it's… Is it enough to… to build this, uh, algorithm. [Qidong Zhou] 15:39:12 Maybe, maybe why you define the cluster, uh… the cluster definition, then, uh, do this calculation if it's, uh… okay or not. For example, add the resolution is okay or not. [Qidong Zhou] 15:39:27 To identify this, for example, this point. At the world for transformation. [Jingzhou Zhao] 15:39:31 Yeah, I would check. [Jingzhou Zhao] 15:39:36 Yeah, I would try putting that for the constant. For… mmm… to combine all the hits further. [Jingzhou Zhao] 15:39:47 Imagine a kid. I think Sidon's question is… from the detector, we cannot read out any ROV And also for the 5, these parameters is… this… all these parameters it to be calculated with… just with, uh. [Jingzhou Zhao] 15:40:07 Educate a note case… uh, head information. [Qidong Zhou] 15:40:11 Yeah, that's definitely the level 1 trigger algorithm for tracking need to be step-by-step. [Qidong Zhou] 15:40:19 So, this trans transformation is one of the steps. We need to evaluate this whole transformation as an input parameter. This information can be get from Which step? And if there's a reconstruction, pre-reconstruction of some, um. [Jingzhou Zhao] 15:40:25 Yes. [Qidong Zhou] 15:40:41 2d tracker, or kind of algorithm is needed before this wave transformation. [Qidong Zhou] 15:40:46 So, that's definitely… we need to define the, uh, as a part of this Level 1 tracking algorithm, uh, design. [Jingzhou Zhao] 15:40:57 The input from the… from the… This Hoff transforms her. [Jingzhou Zhao] 15:41:04 Or the constant information. [Jingzhou Zhao] 15:41:10 Did I use any other… We're tracking everything. [Jingzhou Zhao] 15:41:17 But from detector without heavy information, there are no R… And, uh, certify information. You must calculate Yes. I calculate the… I think I need to get the… So, from the simulation samples. [Jingzhou Zhao] 15:41:38 Do you start from the head? Hit means just… TANOS ID. Tens ID. [Jingzhou Zhao] 15:41:45 And, uh… Because for each channel idea. He also gave me the… precisions, XYZ. Yeah. [Jingzhou Zhao] 15:41:55 You use this XYZ to calculate the R5. Maybe it's just used to… to find the fixed table. [Jingzhou Zhao] 15:42:07 Who translation… Yeah, maybe map all the IDs into… The corresponding privacy. [Jingzhou Zhao] 15:42:19 Just need a very large map. Okay. I think I still… I can still… we can stick that there… SRZ is, like, kidney. [Jingzhou Zhao] 15:42:28 Yeah, so beginning from… I remember really understanding just, uh, estimates the the data rate, uh, there are maybe one layer, there are three. [Jingzhou Zhao] 15:42:41 Pixel will be fired with one head. So, that's the cluster. [Jingzhou Zhao] 15:42:49 Works next. Works next. [Jingzhou Zhao] 15:42:54 Looks like… There's so many issues. [Jingzhou Zhao] 15:42:58 Even for the outermost lane, you can see the physics… physics one. Oh, physics one, yes, there's only one here. [Jingzhou Zhao] 15:43:05 Each, uh, each, uh, each layer will be only one. Yeah, you just assume only when, but from detector view. [Jingzhou Zhao] 15:43:16 Now, there are three hits. Yes, so I need to merge from the architecture. [Jingzhou Zhao] 15:43:21 Into one constant, yes. Yes, that's why I wanted to ask. [Jingzhou Zhao] 15:43:25 Okay. Okay. That's another question or comment? [Jingzhou Zhao] 15:43:33 Uh, one comment for the… Traderie is… maybe you need to… I first defined, uh… cluster. Yes? [Jingzhou Zhao] 15:43:46 And there may be… from the cluster defern, you can find some difference for the… second cluster, and that's, uh… physical… physic class. [Jingzhou Zhao] 15:44:00 Maybe for 2 seconds, uh… maybe it's a huge number. [Jingzhou Zhao] 15:44:06 It's different from the… So it's a physical… his memory. [Jingzhou Zhao] 15:44:14 I don't know if this should study it. [Jingzhou Zhao] 15:44:25 Not counting the background samples, all all the backgrounds and what's, uh… even time. I think it's maybe… most of them, I… by electron… electron or positron? [Jingzhou Zhao] 15:44:43 No, even type. In fact, we're on the human type. Indeed. [Jingzhou Zhao] 15:44:49 So, the samples only. Yeah. Yes. [Jingzhou Zhao] 15:44:56 Ginger's suggestion is… And the EE features, and… maybe it's different cluster size, it's the muu. [Jingzhou Zhao] 15:45:04 Yes, there's, uh, some… some method to… to identifying. [Jingzhou Zhao] 15:45:13 Different types, vertical type. [Jingzhou Zhao] 15:45:22 Uh, so that first is that… from the class. And secondly, use the clusters to reconstruct the track. Next. [Jingzhou Zhao] 15:45:38 Okay. Yeah, I already added a parabolar for this. [Jingzhou Zhao] 15:46:00 Mr. [Jingzhou Zhao] 15:46:06 Don't need… one table for the… I mean, for… for the, uh, results. [Jingzhou Zhao] 15:46:19 Since all your textbook is just the only… It's just a check standalone. Stand-alone, yes. Yeah. Preferably this setup. [Jingzhou Zhao] 15:46:29 Because one punch for each background device, and other, uh. Something like that. Mm-hmm. [Jingzhou Zhao] 15:46:36 How many layers, ethical layers being used for simulation. For the barrier, only 3 layers for the uncap, 4 layer. [Jingzhou Zhao] 15:46:50 But for Vortex, there are two prototypes. Two groups with that. [Jingzhou Zhao] 15:46:55 Why is staging, and one is a pixel. For the Pixel prototype, there may be… craft layers. [Jingzhou Zhao] 15:47:08 Currently, the baseline… the baseline, uh, reference detector is this… Staging. Now, stitching is not so… use the baseline. [Jingzhou Zhao] 15:47:20 Detection to CalConnect. [Jingzhou Zhao] 15:47:26 This is fair. Structure for the Vortex. [Qidong Zhou] 15:47:37 Uh, sorry, Popin. Uh, can you… I have to think about… the 3D whole transformation for your case. [Qidong Zhou] 15:47:47 How you define the file of the tracks? When you… you have two track, for example, in the OTK and ITK, How is the definition of the file? [Jingzhou Zhao] 15:47:59 Hi, sir. Size, stir… send me instances. [Jingzhou Zhao] 15:48:05 Sitting to coordinate Um, from the heat. [Jingzhou Zhao] 15:48:11 Positions to the orange. Are… argument. [Qidong Zhou] 15:48:14 But, um, you have, for example, two… one heater in… on… on the ITK and, uh, one hit on the OTK, then how you define the… buy out these tracks. [Jingzhou Zhao] 15:48:28 Sometime, maybe 10… This part, as example. This is Party, and I assume that the IPs at the… at the zero… Fine. And the files… But that's another hit here. [Jingzhou Zhao] 15:48:42 And the prime is the… Uh, the Apple phone. [Jingzhou Zhao] 15:48:48 Uh, basically to… That's… that's ex… So, horizontal? [Qidong Zhou] 15:48:54 Then you have magnetic field, and how you… the file should be… the IP is on the… in the… in the… Okay, the IEP is 50-0. [Jingzhou Zhao] 15:49:09 Yeah, I can't steal. [Qidong Zhou] 15:49:11 You're… this is not a… You define the every… every check, uh, regarding to the IP, right? [Jingzhou Zhao] 15:49:23 But yes, that's, uh, assumptions, yeah. Uh, some of the ideas that are orange are… All right, true. [Jingzhou Zhao] 15:49:37 If the data is not an origin, then I need to add another co-parents. [Jingzhou Zhao] 15:49:41 Yeah. It's a constant here and another constant here. [Qidong Zhou] 15:49:47 Yes, that's… That's a very… I need to think about this calculation, you can't calculate every heat. [Jingzhou Zhao] 15:49:52 Right. [Qidong Zhou] 15:49:58 Regarding to the… to the… to the IP that will be… have a lot of calculation you have, um… maybe there are several heaters belonging to one of the tracks. [Qidong Zhou] 15:50:12 U… I think we need to think about… we need to first define a kind of a track. [Qidong Zhou] 15:50:19 To connect some of the point as the tracker belonging to one tractor, then calculate this track. [Qidong Zhou] 15:50:27 Over the theta. And, um, calculated zero point. Um, I think maybe even… I will look at how we did this at Beltu, and maybe discuss and think about this again. [Jingzhou Zhao] 15:50:45 It starts off the bat. [Qidong Zhou] 15:50:48 Yeah. [Jingzhou Zhao] 15:50:52 Okay. [Jingzhou Zhao] 15:51:00 So… This hour talk about, and uh… Uh, in the Chinese part, Qidong also suggests to have a discussion at the next meeting. [Jingzhou Zhao] 15:51:13 Uh, to discuss… Uh, however, triggered Uh… software framework. [Jingzhou Zhao] 15:51:23 Uh, we must, uh… uh, study, uh, current, uh. [Jingzhou Zhao] 15:51:28 Cpcsw. Uh, framework to find a way. [Jingzhou Zhao] 15:51:33 How to get, uh… CPC, uh, adaptive, uh. [Jingzhou Zhao] 15:51:39 I have a trigger, uh… software, uh, framework. [Jingzhou Zhao] 15:51:45 Currently, with the, uh, we work for the Journal Experiment. We have a radar framework, but the framework, uh, also can run offline algorithms Uh, for Trunol, but, uh… Uh, we can collect some information. [Jingzhou Zhao] 15:52:07 Uh, to CPC, and uh… Uh, we also invite, uh. [Jingzhou Zhao] 15:52:13 She don't want to give some information from Bell 2. Maybe, Miro, you can give some information? [Jingzhou Zhao] 15:52:21 From our city. Our RCB, so we can discuss, uh. [Jingzhou Zhao] 15:52:28 Framework for the high-level trigger. This, uh, next meeting. [Miroslav Saur] 15:52:33 Sure, I… I can prepare it. Do you want, uh… I guess it would be good to show the whole HRT that's been both stages, HRT1, HRT2. [Jingzhou Zhao] 15:52:34 Topic. [Jingzhou Zhao] 15:52:44 Okay, because… okay, go on. [Miroslav Saur] 15:52:46 And… Yeah, and just thinking it would be good to also invite maybe someone from offline to get some feedback if there would be some restrictions from offline. [Miroslav Saur] 15:53:01 Computing. [Jingzhou Zhao] 15:53:02 Yes, yes. The first stage, we… we just, uh… collects the information together, then we… get some strategies, then we can discuss with the offline guys to find a good way. [Miroslav Saur] 15:53:17 Sounds good [Jingzhou Zhao] 15:53:20 Okay. So, currently, that's… Although, since we're discussion. [Jingzhou Zhao] 15:53:28 And in last, uh… meeting, I just, uh… request the Mira and, uh. [Jingzhou Zhao] 15:53:39 I've come to give some advice for the next R&D. Uh, directions, or… Actions, what do you think about this? [Jingzhou Zhao] 15:53:51 Do you have any suggestions, or…? Comment? [Wolfgang Kuehn] 15:53:54 Well, I think one aspect to look into is what kind of technology should be used for the trigger hardware. I mean, for instance, FPGAs. [Wolfgang Kuehn] 15:54:06 And, of course, you know, the development in this field is very rapid. [Wolfgang Kuehn] 15:54:10 And so, if you use technology which is available now, it might be already outdated, considering the timescale of the CPC project. [Wolfgang Kuehn] 15:54:20 So, I think that's a general… general problem. The other thing is, you know, this very strange political development, so I don't know whether you have heard yesterday. [Wolfgang Kuehn] 15:54:31 The US administration has stopped licensing of cadence and Synopsys software to China. [Wolfgang Kuehn] 15:54:37 So, in principle, that could go on, and maybe they prohibit sales of Xilinx FPGAs to China, so I think one should also have a look into Chinese-based companies. [Wolfgang Kuehn] 15:54:53 I just checked in the web, there are a lot of, uh. [Wolfgang Kuehn] 15:54:55 Also FPGA companies, and there are companies which design electronic, um. [Wolfgang Kuehn] 15:55:02 Design automation software, so I think this is… it's a very strange aspect, but I think one cannot completely ignore it, depending on the developments of the of the next years. [Jingzhou Zhao] 15:55:16 Yeah. [Miroslav Saur] 15:55:17 Same is perfectly valid for GPUs, therefore any usage of GPUs, the situation is Same, or even worse. [Wolfgang Kuehn] 15:55:20 Yeah, exactly, yeah, yeah. Also, uh… Yes, also, also, uh, I mean, uh, optical, optical link technology, all these kind of things could be suddenly stopped. Nobody knows, I mean… This is a completely crazy situation, of course, but I think one cannot… one cannot ignore it, because if you… [Wolfgang Kuehn] 15:55:42 Let's say you base everything on some technology which needs to be imported from the US, and this is suddenly stopped. [Wolfgang Kuehn] 15:55:48 This is a disaster. [Miroslav Saur] 15:55:52 Uh, I could just say on this topic, though. Regarding GPUs, because, okay, I personally think that GPUs are much more flexible, so that's possibly better than FBGAs for, like, more, uh, general processing. [Wolfgang Kuehn] 15:56:07 Yeah, yeah, yeah. [Miroslav Saur] 15:56:07 And I know that I discussed with a couple of IHAP computing people, they are already experimenting with some, uh. [Miroslav Saur] 15:56:14 Uh, with some Chinese GPUs, which are… Not what's, uh, well. [Miroslav Saur] 15:56:20 Another issue with GPUs in general, uh… Is that kind of more and more of our GPUs are aiming AI business. [Miroslav Saur] 15:56:30 Which is perfectly fine, but for us, it's pretty bad, because They are aiming just for a single precision. [Miroslav Saur] 15:56:38 Yeah, that's not good for enough, so really what we need are GPUs which are supporting a scientific computation, floating-point precision. [Miroslav Saur] 15:56:50 And I know that all the IF computing people that are in touch with a couple of Chinese companies which are developing cloud GPUs. [Miroslav Saur] 15:56:58 And they are still providing or planning to provide, uh. Those, uh, those GPUs, which are… early developed for scientific computing, not just for AI, but all this computing, it has… They have a proper floating point-to-point precision. [Miroslav Saur] 15:57:13 And, uh, through pretty fast, even at this high-level precision. So, indeed, it also should be something that, uh. [Miroslav Saur] 15:57:23 Uh, people should start looking online. I can tell you that, okay, I got in touch with them as well, because, uh. [Miroslav Saur] 15:57:30 Was idea that we could tie to… to benchmark some of those scouts, also for LHCB. [Miroslav Saur] 15:57:36 And to see how they… well, how they are compared to standard NVIDIA cards. [Miroslav Saur] 15:57:43 Yeah, but as Volvgan said, all these things, they have to be taken into account. [Miroslav Saur] 15:57:50 And even, let's say… mostly we would be aiming for some specific hardware, with some specific properties. [Miroslav Saur] 15:57:57 There always must be some fallback, uh, solution which would not, uh. [Miroslav Saur] 15:58:03 Main… well, which still would be feasible, and it still would be good enough to… to support the whole thingy system. [Jingzhou Zhao] 15:58:14 Okay, so for this issues… We also are synced before, because In our design, uh, design view, we… also try to design our common bot with a dynamic SPJ, uh, it's our design. [Jingzhou Zhao] 15:58:35 In our design, but for the GPU, Uh, in different projects, we also… start to study the domestic, uh, GPUs. [Jingzhou Zhao] 15:58:47 Devices, uh, the issues there, we… We must define some, uh… Femistic. [Jingzhou Zhao] 15:58:57 Vendors or devices to… to deal with these issues. [Miroslav Saur] 15:59:05 Well, personally, I would say because there are a couple of those companies, uh, but at this moment, uh. [Miroslav Saur] 15:59:12 Most of the GPUs are not really tested for scientific computation of our kind, yeah, so… I would say at the beginning would be really quite, quite nice, uh, to… to do a benchmarking, yeah, because of those cards, which is, of course, complicated, because many of them, they are using very different, uh. [Miroslav Saur] 15:59:31 Uh, development, well… the programming language, the whole… Sorry. The whole way, how to program those GPUs are actively different, because they… that are using different systems, so… Uh, but still… that can be overcome, yeah, one can use some more generic, uh… [Miroslav Saur] 15:59:55 Uh, well, there's this open seal and so on, which I think it could support all those cards. [Miroslav Saur] 16:00:02 So that could worry about the, uh, yeah, back to my main point. [Miroslav Saur] 16:00:06 Would be quite nice, maybe, to get in touch with a couple of those companies, to have a couple of cards, and then retail benchmarking. [Miroslav Saur] 16:00:14 Then just directly from the beginning, say that, okay, we are going just with, uh. [Miroslav Saur] 16:00:18 One, uh, one solution. [Jingzhou Zhao] 16:00:22 Okay, okay, fine. And, uh… Uh, recently, I also… got some, uh, information. [Jingzhou Zhao] 16:00:32 From our computer center. Uh, there are some companies who provide GPU, uh, for tests, you can try it to… build some bench… and, uh… Benchmark test the bed to… to know whether this Chinese company's paradactic could be… whether it could be… meet our requirement. [Jingzhou Zhao] 16:01:04 But we must do that. Trying to find some… uh, usable… devices, both for IPG and GPU. [Miroslav Saur] 16:01:18 Just to say about making a good benchmark, or be able to properly benchmark some cards, it's not avoid, let's say, for one week. [Miroslav Saur] 16:01:27 Because what one would need, at least on, let's say, basic… Uh, tracking algorithm to point that uh, tracking algorithm to… to those cards. That's been also some kind of input-output, especially input… or just to generate things. [Miroslav Saur] 16:01:48 And as I said, because those cars, they are using quite different development framework. [Miroslav Saur] 16:01:54 It may not be so simple, so… It would not be, like, one week, but maybe… I don't know if, uh, from LHCB experience, I remember people were saying, MAND cards. [Miroslav Saur] 16:02:07 It took, like, 2 months. To really have something what could be counted as a prop, uh… benchmark, and those were people which were already familiar with, uh, with all the code. [Miroslav Saur] 16:02:19 Was originally written on LHCB. This is just to give you some idea, so… This is benchmarking, for sure, has to be done. [Miroslav Saur] 16:02:26 But it needs to be started well in advance. Because it will not be done, uh, within one week. [Jingzhou Zhao] 16:02:36 Yes, Mr. Rhymes. But, uh… That's before the benchmark, we must prepare us some tactical algorithms. [Jingzhou Zhao] 16:02:47 Including algorithm. To try to… to cast the old… all this… all these, uh… possible devices. [Miroslav Saur] 16:03:03 Yes, sure. We should have some, let's say, baseline, mostly in terms of logic, how, let's say, how to do some of the decoding, clustering, and so on. [Miroslav Saur] 16:03:14 And then, uh, so this would be this baseline will have to be ported to those… the different, uh… different cards based, maybe they will need some slightly different way how to work, uh. [Miroslav Saur] 16:03:28 With the code. But for exactly as you said, we need to have this, this basic version. [Miroslav Saur] 16:03:35 Bidrogen can be used elsewhere. [Jingzhou Zhao] 16:03:37 Yeah, yes, we need more. More guys to… to join us to… step-by-step. [Jingzhou Zhao] 16:03:46 To try this. [Jingzhou Zhao] 16:03:54 Just Wolfgang said. Yesterday… America, Turbo. [Jingzhou Zhao] 16:04:01 Yeah, forbidding to use the Canon software for the license for China, so… We must try some way into. [Wolfgang Kuehn] 16:04:10 Now, Cadence and synopsis, both. [Jingzhou Zhao] 16:04:15 And a sib… So it says, synopsis. [Jingzhou Zhao] 16:04:19 Do we… I just want to point out that there are also some Chinese companies working on the software tools. [Jingzhou Zhao] 16:04:30 I heard that he mentioned that also, huh? Am I right, Wolfka? [Wolfgang Kuehn] 16:04:35 Pardon? I didn't get this? [Jingzhou Zhao] 16:04:37 I thought that you mentioned that there are also some Chinese companies who are working on the tools. [Wolfgang Kuehn] 16:04:44 Yes, there is… I have a list here. Empyrean Technology and Expedic Technology, and Primarius Technology. They are… they are developing these tools, but however. [Wolfgang Kuehn] 16:04:54 Of course, you have to consider that, I mean, in particular for ASIC development, cadence and Synopsys are really very, very important, huh? [Wolfgang Kuehn] 16:05:03 Maybe not so much for FPGA design, but for ASIC development, so this is more the front-end part of the detectors, which are affected, and I mean, of course, this all has to integrate then with the companies producing, actually, the chips and so on. This is, uh… [Wolfgang Kuehn] 16:05:23 It's a very complex, uh, complex, uh, thing there. [Jingzhou Zhao] 16:05:31 Yeah, there's… for us, there's no choice, huh? [Wolfgang Kuehn] 16:05:35 Yeah, sure, I think one has to really look into this, and I mean, at least, maybe in a few years, this is all over, this nonsense, but you cannot count in it, and so there has to be… one has to really make an alternative. [Wolfgang Kuehn] 16:05:50 Based on domestic products, I think. Otherwise, that could be really a showstopper. [Jingzhou Zhao] 16:05:57 Yep. Yeah, yes. [Jingzhou Zhao] 16:06:07 Okay, so we're almost done. Yeah. [Jingzhou Zhao] 16:06:16 Themistic Happy deal, or something? [Qidong Zhou] 16:06:25 Uh, for now, uh… current TMs still using settings as PJs, but I think it's necessary to think about some demonstration on other a PG product, uh, inside China. [Jingzhou Zhao] 16:06:40 Okay, so that's our… [Qidong Zhou] 16:06:41 Oh, I… [Wolfgang Kuehn] 16:06:41 You have to also think about the networking components. I mean, high-speed optical links, and so on, this is all coming from whatever, Avago, Broadcom, and these companies Uh, which could also be part of an embargo. So when… I think one has to look at the whole. [Wolfgang Kuehn] 16:06:58 It infrastructure, which is needed. [Jingzhou Zhao] 16:07:01 Yeah, yes. [Qidong Zhou] 16:07:02 But for that part, I heard from my colleagues inside our university that Bali had provided also Um, quite the kind of good module that could be used for high-speed, um. [Qidong Zhou] 16:07:18 Transformation, uh, at, um, um… I just seem to… Um, we have been used, for example, our goal or some other products that we have been used. [Qidong Zhou] 16:07:30 But for the FPGs, um, for the advanced, um… A much more advanced, um… capability of the FPG, for example, some EA engine or… heterogeneous computing platform that has been provided by Cynics. Currently, I think. [Wolfgang Kuehn] 16:07:31 Yes [Qidong Zhou] 16:07:50 Inside China, we don't have such kind of, um… Such kind of chips, but… I think we need to also still using, uh, traditional kind of FPGs. [Qidong Zhou] 16:08:01 Like, um, um… vertex… to a scale kind of class. [Qidong Zhou] 16:08:07 Fpgs inside China that can be produced by some China companies. [Qidong Zhou] 16:08:18 But I have never used this kind of FPH chips for designing, including the software, uh, and the chips, so… I think maybe you can go, um, expressed, we can maybe… think about to buy some prototype also. [Qidong Zhou] 16:08:38 Tests on both… both kind of platform for this… Um, however, designing our algorithm designing point of view. [Jingzhou Zhao] 16:08:51 We must do that. Yeah. [Jingzhou Zhao] 16:08:59 Okay… So, we go back to the cave. [Jingzhou Zhao] 16:09:07 Stay our life. [Jingzhou Zhao] 16:09:13 So, any other… Question, comment, or suggestion? [Jingzhou Zhao] 16:09:22 Almost to finish the TDR, but we… the real stats real RND, very, very… So… [Jingzhou Zhao] 16:09:37 Okay, if not, we can stop here, thank all of you, join. [Jingzhou Zhao] 16:09:42 The meeting? Okay? See you next time! [Qidong Zhou] 16:09:50 Bye. [Wolfgang Kuehn] 16:09:50 So, Jen? [Hongyu Zhang] 16:09:50 I see you. [Jingzhou Zhao] 16:09:51 Bye-bye. So. [Jingzhou Zhao] 16:09:59 An advertising, like, do you know what? To the shipping that you don't like that. [Jingzhou Zhao] 16:10:11 Um.