Fixing GPU & CPU Benchmarks: Introducing Animation Error Engineering Discussion
video description
Date: 2024-03-13
Related videos
Comments and reviews: 19
gsestream
denoising is not required for ray traced stuff, use raster ray traced method, with separate lighting step, also rasterized collection, key word uniform complete sampling not random sparse sampling. like 32x32x6 cubemaps = 6000 samples per pixel per triangle. yes one light map value per triangle (assumed the more triangles the smaller the triangles the more accurate per triangle lighting). fix your de-noising ray-tracing algorithm, by removing them, use raster algorithms and filters for ray tracing. well frames per second and frame time are different measures. also if ray traced output is not stable for identical starting conditions and same in shape as raster, its wrong. also when lighting is completely separated and asynch from the frame raster collection rays, then you cant actually say what is rendering time. if you only render first rays or only reflection/refraction rays from camera, you know exactly how many rays per frame you need to calculate. but those dont do lighting, only camera ray collection. per surface pixel/triangle light accumulation is done separately. yes real-time light-map baking with rays/raster, and the current state of the lighting is rendered by the primary camera rays. you can use raster hardware for most real-time baking, and ray tracing hardware for camera ray raster, and ai/ml cores for the raster smoothing/upscaling style calculations.
reply
denoising is not required for ray traced stuff, use raster ray traced method, with separate lighting step, also rasterized collection, key word uniform complete sampling not random sparse sampling. like 32x32x6 cubemaps = 6000 samples per pixel per triangle. yes one light map value per triangle (assumed the more triangles the smaller the triangles the more accurate per triangle lighting). fix your de-noising ray-tracing algorithm, by removing them, use raster algorithms and filters for ray tracing. well frames per second and frame time are different measures. also if ray traced output is not stable for identical starting conditions and same in shape as raster, its wrong. also when lighting is completely separated and asynch from the frame raster collection rays, then you cant actually say what is rendering time. if you only render first rays or only reflection/refraction rays from camera, you know exactly how many rays per frame you need to calculate. but those dont do lighting, only camera ray collection. per surface pixel/triangle light accumulation is done separately. yes real-time light-map baking with rays/raster, and the current state of the lighting is rendered by the primary camera rays. you can use raster hardware for most real-time baking, and ray tracing hardware for camera ray raster, and ai/ml cores for the raster smoothing/upscaling style calculations.
reply
DelticEngine
GamersNexus Love these videos! There has obviously been a lot of detailed development and research going on. A tool such as PresentMon could also be useful in comparing different driver releases for GPUs and other devices used while gaming.
I have two questions:
1) Will PresentMon compile and run on a Linux system I want to be able to compare the performance between Linux and Windows running on the same machine. Linux also has low-latency kernels available, unlike Windows, and it would be interesting to see what benefits, if indeed any, one of these kernels gains.
2) Will PresentMon work with multiple GPUs I would like to know what advantages there are, and any disadvantages, with such a configuration. Running multiple GPUs is of interest because of the cost difference between having a single flagship GPU or two (or more) lower GPUs.
reply
GamersNexus Love these videos! There has obviously been a lot of detailed development and research going on. A tool such as PresentMon could also be useful in comparing different driver releases for GPUs and other devices used while gaming.
I have two questions:
1) Will PresentMon compile and run on a Linux system I want to be able to compare the performance between Linux and Windows running on the same machine. Linux also has low-latency kernels available, unlike Windows, and it would be interesting to see what benefits, if indeed any, one of these kernels gains.
2) Will PresentMon work with multiple GPUs I would like to know what advantages there are, and any disadvantages, with such a configuration. Running multiple GPUs is of interest because of the cost difference between having a single flagship GPU or two (or more) lower GPUs.
reply
paulboyce8537
The point is that INTEL ARC gives lot better results with lot less FPS in current testing. Paper values do not match the experience. AMD/Nvidia are close the same architecture. INTEL is different architecture and for measuring you need new ways that apply to all. This has been my concern for a long time. Now we are getting a more in depth testing measure. I don't think this is just to make ARC to look better. I think it is more about taking in consideration the difference in architecture. I see this because Battlemage is getting very close now and if tested the old way it will look bad. This will also lift the Alchemist. Lot happening and to bring out better ways to test is very good for the consumer.
reply
The point is that INTEL ARC gives lot better results with lot less FPS in current testing. Paper values do not match the experience. AMD/Nvidia are close the same architecture. INTEL is different architecture and for measuring you need new ways that apply to all. This has been my concern for a long time. Now we are getting a more in depth testing measure. I don't think this is just to make ARC to look better. I think it is more about taking in consideration the difference in architecture. I see this because Battlemage is getting very close now and if tested the old way it will look bad. This will also lift the Alchemist. Lot happening and to bring out better ways to test is very good for the consumer.
reply
jinkun02
1% lows have been bothering me for a long bit, because it's not the low framerate represented in that number that I necessarily experience/notice, it's the flat-out halts, the jitters, the massive chunks of redrawn frames. It's hard to parse, what you REALLY want to know his how often am I going to notice I do not have a new frame ready, that's what you FEEL. I know it's hard to represent these sort of things with numbers and averages. By design they don't really illustrate the extremes very well. Software input to frame data is gonna be sick as well!
reply
1% lows have been bothering me for a long bit, because it's not the low framerate represented in that number that I necessarily experience/notice, it's the flat-out halts, the jitters, the massive chunks of redrawn frames. It's hard to parse, what you REALLY want to know his how often am I going to notice I do not have a new frame ready, that's what you FEEL. I know it's hard to represent these sort of things with numbers and averages. By design they don't really illustrate the extremes very well. Software input to frame data is gonna be sick as well!
reply
shuckle2
I don't know if it's related to Animation Time Error, and I don't know if it's a thing that happens in almost every game, but I remember when playing games without enough VRAM or with very low inconsistent FPS, sometimes the games didn't just feel low FPS, they didn't just feel stuttery, but the animations of the characters felt jumpy (even worse with the physics), it seemed like the body coordinates of their animations moved inaccurately and like online rubberbanding. Kinda like PS1 polygon wobbling but based on movement.
reply
I don't know if it's related to Animation Time Error, and I don't know if it's a thing that happens in almost every game, but I remember when playing games without enough VRAM or with very low inconsistent FPS, sometimes the games didn't just feel low FPS, they didn't just feel stuttery, but the animations of the characters felt jumpy (even worse with the physics), it seemed like the body coordinates of their animations moved inaccurately and like online rubberbanding. Kinda like PS1 polygon wobbling but based on movement.
reply
rd-um4sp
average seems like a bad statistic for FPS from the start. Maybe a median or mode (with some tweaking) would be more representative of the player experience. Some variance or standard deviation too. But then again, you'd have to explain statistics concepts to the general audience every single video that would tank the views.
Oh, that has nothing to to with the video discussion Sorry about that. I'll try to do better next time.
reply
average seems like a bad statistic for FPS from the start. Maybe a median or mode (with some tweaking) would be more representative of the player experience. Some variance or standard deviation too. But then again, you'd have to explain statistics concepts to the general audience every single video that would tank the views.
Oh, that has nothing to to with the video discussion Sorry about that. I'll try to do better next time.
reply
robertlawrence9000
It's so good to see Intel getting more into graphics and gameplay performance. I really hope they keep striving to compete. Seeing early MSI Claw performance tests, it's not looking so good against the competition. I really hope they can fix that and keep getting way better each generation. I know they can do it! They just got to go all out!
reply
It's so good to see Intel getting more into graphics and gameplay performance. I really hope they keep striving to compete. Seeing early MSI Claw performance tests, it's not looking so good against the competition. I really hope they can fix that and keep getting way better each generation. I know they can do it! They just got to go all out!
reply
julianB93
Thats why i love you. Now we get frame times and stuttering as a category in gpu reviews. It blows my mind how bad some games run on certain gpus. In my case ist an amd gpu and it just got better after going for shaders always on in the registry. Also i saw insane performance in games that pre cache its shaders like Lies of P as an example.
reply
Thats why i love you. Now we get frame times and stuttering as a category in gpu reviews. It blows my mind how bad some games run on certain gpus. In my case ist an amd gpu and it just got better after going for shaders always on in the registry. Also i saw insane performance in games that pre cache its shaders like Lies of P as an example.
reply
MikeNovelli
Why does Presentmon not work in some applications Like Furmark... I have my own workaround but I would prefer to just get the overlay right on top..
Why does Presentmon give us so many options for telemetry and then limit the amount of graphs you can see on the overlay in the way it does
The UI/UX needs some changes
reply
Why does Presentmon not work in some applications Like Furmark... I have my own workaround but I would prefer to just get the overlay right on top..
Why does Presentmon give us so many options for telemetry and then limit the amount of graphs you can see on the overlay in the way it does
The UI/UX needs some changes
reply
902384902384
Put the display on a web server so I can check it on my phone instead of an overlay. (Pipe data from libre hardware monitor too so I can have anything else I might want).
Set the default color for CPU/GPU graph lines to company brand colors for faster recognition of what you are seeing.
Looks great. Thx.
reply
Put the display on a web server so I can check it on my phone instead of an overlay. (Pipe data from libre hardware monitor too so I can have anything else I might want).
Set the default color for CPU/GPU graph lines to company brand colors for faster recognition of what you are seeing.
Looks great. Thx.
reply
Modna89
If these were the only videos you had I would be content. Interviewing these highly technical guys is so entertaining. Plz do all you can (P.S.: not to say your other content isn't great, it's just that this stuff is SO good)
reply
If these were the only videos you had I would be content. Interviewing these highly technical guys is so entertaining. Plz do all you can (P.S.: not to say your other content isn't great, it's just that this stuff is SO good)
reply
9SMTM6
When they came up with that weird smooth sync I thought that is what they're wasting time on
But these new metrics are genuine improvements for the whole industry, and at least I will remember that and be thankful to Intel.
reply
When they came up with that weird smooth sync I thought that is what they're wasting time on
But these new metrics are genuine improvements for the whole industry, and at least I will remember that and be thankful to Intel.
reply
kamipls6790
This guy keeps Intel's reputation up. Love to see him in your vids. I actually own an Arc 770 16gb limited edition
It has it's use cases and I'm happy to own one!
I hope they keep their head up and go on improving
reply
This guy keeps Intel's reputation up. Love to see him in your vids. I actually own an Arc 770 16gb limited edition
It has it's use cases and I'm happy to own one!
I hope they keep their head up and go on improving
reply
LegendaryKira
You may want to consider a camera stand or something when the majority of time you are only in 1 location. The jidder can be very odd (motion sickness inducing for me) when it is in full view frame shot. Great video!
reply
You may want to consider a camera stand or something when the majority of time you are only in 1 location. The jidder can be very odd (motion sickness inducing for me) when it is in full view frame shot. Great video!
reply
TexasJoe1985
No joke, I started watching this video and my machine started randomly dropping frames every two seconds. Closed every open app, and it was still happening. Rebooted and the problem is gone. I blame this video.
reply
No joke, I started watching this video and my machine started randomly dropping frames every two seconds. Closed every open app, and it was still happening. Rebooted and the problem is gone. I blame this video.
reply
MaxDad7
I’m hoping there will be a Linux version as this would be a nice-to-have. If not, then I hope the project is truly open source so the Linux community might be able incorporate it. Thanks for the presentation Tom!
reply
I’m hoping there will be a Linux version as this would be a nice-to-have. If not, then I hope the project is truly open source so the Linux community might be able incorporate it. Thanks for the presentation Tom!
reply
marekciostek1458
This all 1% low, 0.1% low, not stable frame pacing, drops from 100fps to under 60fps drive gamers mad. This is why locked (V-synch, free-sync, G-sync, etc.) 60, 75, 90 (...) consistet FPS is better.
reply
This all 1% low, 0.1% low, not stable frame pacing, drops from 100fps to under 60fps drive gamers mad. This is why locked (V-synch, free-sync, G-sync, etc.) 60, 75, 90 (...) consistet FPS is better.
reply
AlexSchendel
As an Intel engineer, GO TAP!
I always enjoy seeing TAP talking about Intel Arc development. Though I do not work on Intel Arc, so it makes me wish TAP could present things on other projects haha.
reply
As an Intel engineer, GO TAP!
I always enjoy seeing TAP talking about Intel Arc development. Though I do not work on Intel Arc, so it makes me wish TAP could present things on other projects haha.
reply
ozzieggg
Hope they fix the idle power draw. I was going to buy an arc card for the home server because of the very nice video encoder, but the idle power draw is just unacceptable for that use case.
reply
Hope they fix the idle power draw. I was going to buy an arc card for the home server because of the very nice video encoder, but the idle power draw is just unacceptable for that use case.
reply
Add a review, comment
Other channel videos
