My favorite hat so far 😉 I like this format. It would have been nice if you could make more visual representations of some concepts/architectures that you present in this video. Thank you for sharing doc, see you in the next one !
This hat is the winner to be honest. I finally figured out how to wear it properly. :) Yeah I realized I should at least have put the NVidia stock price chart. And a double-buffer diagram. I'll focus more on the visual representations in the future. See you on the next one!
WOW, this is became my top favorite computer tech channel on YT, great topic and great content, no carb talk and non sense blabbing as other many channels around the internet ..
@@DrWaku well, you nailed it dr. waku, you really really need to do some sort of campaign, so you can reach more audience with this neat content, and believe me u r a top notch in ur field, take my word, iam 50y man working in tech industry since the 8088 and green monitors.
I love how one branch of science so often builds upon tools developed for other branches of science. Makes everything seem so fundamentally interconnected.
I liked this video. I did get more of a general idea of what makes GPUs do what they do. Of course, if I told you that now I understand them completely that would be a lie. I don't
Did you just say, that if I write a shader that loops a 10 pixel by 10 pixel array with two if statements, it will predictably fire 400 threads per pixel to start my 6 or 7 line program?! Yes you did. That is spooky. And who decides this? The microcode in the GPU i guess, with reference to the number of instructions and the known delay time of all pipelines.
I haven't watched the video yet (I'll watch it later 😅), but it really annoys me how NVIDIA calls SIMD lanes "cores". It's basically disinformation as far as I'm concerned!
I explain more background in this video than normal. It's also much longer as a result. Let me know how you like this format.
Not as many people are clicking on it, but I'm not sure if that's because of the length or the subject matter.
@@DrWakuBe patient. Long form builds slower; many save it for later, for example. It's a great video. Thanks for sharing.
I thought it was good.
@@K.F-R Didn't think of people saving for later, but it makes sense! Thanks!
I've enjoyed it a lot, it's pretty well explained. Thanks for the content!
This is the third video of yours I've watched. You struck a balance between detail and ease of understanding. Congratulations.
Thank you very much! I hope to see you on more videos in the future!
This video is at a good level, something for everyone without getting too deep into details. And length was good too. Thanks
Thanks for your comment! I was worried since it was over 20min :/ See you in the next one
My favorite hat so far 😉
I like this format.
It would have been nice if you could make more visual representations of some concepts/architectures that you present in this video.
Thank you for sharing doc, see you in the next one !
This hat is the winner to be honest. I finally figured out how to wear it properly. :)
Yeah I realized I should at least have put the NVidia stock price chart. And a double-buffer diagram. I'll focus more on the visual representations in the future. See you on the next one!
WOW, this is became my top favorite computer tech channel on YT, great topic and great content, no carb talk and non sense blabbing as other many channels around the internet ..
I'm honored!
@@DrWaku well, you nailed it dr. waku, you really really need to do some sort of campaign, so you can reach more audience with this neat content, and believe me u r a top notch in ur field, take my word, iam 50y man working in tech industry since the 8088 and green monitors.
This was very informative. Thanks!
Very good and detailed on the technical side, not yet fully watched but still learned a lot.
💥 One of my favorite all time channels on UA-cam. Here we are in good hands of the real knowledge, to learn more and more ! 🎉❤
Thank you very much for your kind comment! :) see you around!
Love your videos, can't get enough. Thanks.
Thanks for the comment ;)
I love how one branch of science so often builds upon tools developed for other branches of science. Makes everything seem so fundamentally interconnected.
This is your second video I watched and loved it! Thank you
Love it when you leave lots of comments, I can track your interest through my content :)
Awesome dissertation, thank you.
It's refreshing to see a tech channel where the presenter is still in front of the camera.
очень недооценённый канал! отличные видео! Спасибо 🙏 very underestimated Channel, Great videos, thanks a lot!
I liked this video. I did get more of a general idea of what makes GPUs do what they do. Of course, if I told you that now I understand them completely that would be a lie. I don't
I really enjoyed this video.
Thanks for commenting! :)
You were into GPUs before it was cool to be into GPUs 😎
Yep haha. Early to the trend for once
Fascinating!
Thanks for watching!
Thank you for always putting out quality content! Wish you well!! @@DrWaku
FYI the video chapters have disappeared. Have you changed anything in the description?
I changed the title a few times, maybe that did it. I'll look into it. Apologies
I hit save a few times, didn't work, but when I deleted the chapters and then re-added, they came back! Strange. Thanks for pointing it out.
@@DrWaku I'm happy you were able to diagnose it! I remember fumbling with the description before and not knowing what I did to fix it
@@DrWaku no need to apologize at all. Just looking out 😊
Did you just say, that if I write a shader that loops a 10 pixel by 10 pixel array with two if statements, it will predictably fire 400 threads per pixel to start my 6 or 7 line program?! Yes you did. That is spooky. And who decides this? The microcode in the GPU i guess, with reference to the number of instructions and the known delay time of all pipelines.
Btw, even Americans use Celsius for computer temperatures.
I haven't watched the video yet (I'll watch it later 😅), but it really annoys me how NVIDIA calls SIMD lanes "cores". It's basically disinformation as far as I'm concerned!
You didn't mention A100 vs H100.
True, it's more about the history and hardware suitability. I'll make sure to talk about state of the art sometime too
Do those gloves work.
Yes, for me. ua-cam.com/video/S8AhH2UDtj8/v-deo.html
Terrrific explainer. People actually need to understand the whole history to get why things are the way they are in GPU’s. Appreciate it a lot.
Thank you! I agree the history of this hardware seems to always be overlooked but it has important implications