3 ms·

Measured as one percent lows and again, most GPUs under most Conditions simply don't saturate PCIe enough for that to make a difference. Der8auer and GamersNexu

by dtx1 3y ago

Measured as one percent lows and again, most GPUs under most Conditions simply don't saturate PCIe enough for that to make a difference. Der8auer and GamersNexus on youtube have plenty of examples.

Retric 3y ago
1% lows in a given benchmark isn’t really measuring stutter in practice, what you want is percentage of frames that don’t finish in time. If once a minute you miss 4 frames in a row that’s noticeable even if everything else is rock solid. The thing is people adjust their resolution/settings to reach acceptable FPS, thus it’s fairly GPU independent. What matters is rendering volatility as assets are loaded etc.
- sudosysgen 3y ago
  That metric is irrelevant now that we have adaptive refresh rates. You don't really "miss" a frame anymore. 1% low is now the correct metric.
  - Retric 3y ago
    If the pipeline stalls for 70ms it’s a problem even if in theory you don’t “miss” a frame you still aren’t getting new frames until rendering finishes. All adaptive refresh rates do is avoiding unwanted delays after the frame finishes.
    - sudosysgen 3y ago
      That's not how it works. A 70ms pipeline stall means a ~90ms frame which means that your 1% framerate low is going to be reduced. Adaptive refresh rate means that only frame rate matters. You can't miss a frame, you can only have a frame take longer, and that's measured perfectly well by 1% or 0.1% low frame rates. Also I don't see why an M series CPU is going to be any better at feeding the (asynchrous, multiple frames in flight) GPU rendering pipeline than a modern x86 CPU, when both are just as fast. macOS is also pretty bad at realtime scheduling for heavy workloads compared to modern Linux and Windows.
      - Retric 3y ago
        Yes occasional 90ms stall reduces 1% frame rates, but less than how annoying it is. A 10 minute test at 120 FPS is 10 * 60 * 120 ~= 72,000 frames. Your 1% lows is an average of ~720 frames but so what if of 300 them are at 60 FPS you’re going to have trouble noticing. While if you’re almost rock solid at 120FPS a dozen 70ms stalls it’s really obvious that something is wrong but the calculated 1% low’s may actually look better. This is especially true as those stalls generally correlate with interning things happening. The M series CPU comparison people are talking about isn’t average FPS or even 1% lows where dedicated graphics cars have an advantage, the comparison is what happens when things go very wrong. And it’s in that very specific case when they may have an advantage.
        wtallis 3y ago
        > While if you’re almost rock solid at 120FPS a dozen 70ms stalls it’s really obvious that something is wrong but the calculated 1% low’s may actually look better. I think this is an instance of the coordinated omission problem: you cannot simply measure the latency of the frames that were completed, but instead have to consider the frames that should have been delivered during a stall. Looking at frame time percentiles means a stall only penalizes your metric with one bad frame, when it should be penalized for missing several frames and showing the user an increasingly stale image for several average-frametimes.
        sudosysgen 3y ago
        This isn't actually true. Looking at aveeages will in fact reflect correctly on average frame staleness over a sampling period. A single frame that's 5x too long or 4 frames that are 2x too long have the same impact on the average delta of frame staleness on the total sum. Think of it like Little's law in queing theory , it's better to simply just look at different percentiles, like 99 or 99.9th. The average framerate correctly indicates the average staleness.
        wtallis 3y ago
        > A single frame that's 5x too long or 4 frames that are 2x too long have the same impact on the average delta of frame staleness on the total sum. It's not clear to me: are you saying those two scenarios should be quantified as equally bad? Because it seems pretty obvious to me that the 5x outlier is qualitatively much worse.

by dtx1 3y ago

Measured as one percent lows and again, most GPUs under most Conditions simply don't saturate PCIe enough for that to make a difference. Der8auer and GamersNexus on youtube have plenty of examples.

Retric 3y ago

1% lows in a given benchmark isn’t really measuring stutter in practice, what you want is percentage of frames that don’t finish in time. If once a minute you miss 4 frames in a row that’s noticeable even if everything else is rock solid. The thing is people adjust their resolution/settings to reach acceptable FPS, thus it’s fairly GPU independent. What matters is rendering volatility as assets are loaded etc.

sudosysgen 3y ago
That metric is irrelevant now that we have adaptive refresh rates. You don't really "miss" a frame anymore. 1% low is now the correct metric.
- Retric 3y ago
  If the pipeline stalls for 70ms it’s a problem even if in theory you don’t “miss” a frame you still aren’t getting new frames until rendering finishes. All adaptive refresh rates do is avoiding unwanted delays after the frame finishes.
  - sudosysgen 3y ago
    That's not how it works. A 70ms pipeline stall means a ~90ms frame which means that your 1% framerate low is going to be reduced. Adaptive refresh rate means that only frame rate matters. You can't miss a frame, you can only have a frame take longer, and that's measured perfectly well by 1% or 0.1% low frame rates. Also I don't see why an M series CPU is going to be any better at feeding the (asynchrous, multiple frames in flight) GPU rendering pipeline than a modern x86 CPU, when both are just as fast. macOS is also pretty bad at realtime scheduling for heavy workloads compared to modern Linux and Windows.
    - Retric 3y ago
      Yes occasional 90ms stall reduces 1% frame rates, but less than how annoying it is. A 10 minute test at 120 FPS is 10 * 60 * 120 ~= 72,000 frames. Your 1% lows is an average of ~720 frames but so what if of 300 them are at 60 FPS you’re going to have trouble noticing. While if you’re almost rock solid at 120FPS a dozen 70ms stalls it’s really obvious that something is wrong but the calculated 1% low’s may actually look better. This is especially true as those stalls generally correlate with interning things happening. The M series CPU comparison people are talking about isn’t average FPS or even 1% lows where dedicated graphics cars have an advantage, the comparison is what happens when things go very wrong. And it’s in that very specific case when they may have an advantage.
      - wtallis 3y ago
        > While if you’re almost rock solid at 120FPS a dozen 70ms stalls it’s really obvious that something is wrong but the calculated 1% low’s may actually look better. I think this is an instance of the coordinated omission problem: you cannot simply measure the latency of the frames that were completed, but instead have to consider the frames that should have been delivered during a stall. Looking at frame time percentiles means a stall only penalizes your metric with one bad frame, when it should be penalized for missing several frames and showing the user an increasingly stale image for several average-frametimes.
        sudosysgen 3y ago
        This isn't actually true. Looking at aveeages will in fact reflect correctly on average frame staleness over a sampling period. A single frame that's 5x too long or 4 frames that are 2x too long have the same impact on the average delta of frame staleness on the total sum. Think of it like Little's law in queing theory , it's better to simply just look at different percentiles, like 99 or 99.9th. The average framerate correctly indicates the average staleness.
        wtallis 3y ago
        > A single frame that's 5x too long or 4 frames that are 2x too long have the same impact on the average delta of frame staleness on the total sum. It's not clear to me: are you saying those two scenarios should be quantified as equally bad? Because it seems pretty obvious to me that the 5x outlier is qualitatively much worse.