Recent content by Scraph

That might've been their attempt to compile a non-SIMD version since the compiler will make some attempt to use SIMD on its own. Of course, SIMD instruction sets can be disabled explicitly through compiler flags, while still producing an otherwise optimized build. But even if that comparison...

Reasonable question. If you only render the frames you need, you also know when they'll display. And if you know when they'll display, then you can render frames to show the right thing at the right time. In other words, you can _eliminate_ latency this way. To avoid missing the refresh, you...

The stock 9800X3D trails the 7800X3D, 7950X3D, 14900K, 13900K, and barely beats (<1 fps difference) the 14700K, 13700K, and Ultra 5 245K for 4K minimum frame rate. https://www.techpowerup.com/review/amd-ryzen-7-9800x3d/21.html Have you ever heard of minimum frame rate?

That would be pretty simple to assume. Too simple. CPU performance doesn't scale linearly once you start implicating cache and memory. Remember the X3D part in the 9800X3D name? The cache is why X3D is such a big deal. So, it's also a big deal when a benchmark interacts differently with the...

And they are consumers, not developers, and don't understand that the same could be achieved without all that overkill. A game can render one frame right before each refresh and also get low latency. It's a simple problem with an equally simple solution.

And it will. AMD X3D is king of "1080p low" benchmarks. But, no one actually plays games at "1080p low". No one (actually) plays games at a faster framerate than their monitor, either. If your CPU can give you 60fps solid (or w/e your monitor is), you're good. Your monitor won't show you...

Of course, there are different aspects to performance and as many benchmarks to measure them. Take the 285K, for example. It's deemed a "failure" because it falls short of competition in gaming benchmarks. But in professional usage, it's on top; compiling software, Cinebench rendering, etc...

To be clear, intrinsics certainly do cover the vfmadd231ps instruction. The "_mm512_fmadd_ps" covers it along with the 132 and 213 variants. The difference in the 3 isn't relevant at the C abstraction level, so there would be no sense in distinguishing between them there. Likewise, whether the...

Note that C => SSSE3 is where the big performance gain occurs. From there, SSSE3 => AVX512 gives only another ~2x improvement. Their results show that SIMD is much faster than no-SIMD for their application. But their performance margin between SSSE3 and AVX512? That's "meh" and basically...

I'm doubtful that hand-written assembly would've produced much better code than compiler instrinsics. Compiler intrinsics quite often have a 1-to-1 correspondence to SIMD instructions anyways. Where they don't, there are "reasons" (emulation, abstraction, or optimization). As important for SIMD...

"However, the screenshot also indicates the AMD RX 6950 XT is being pushed to 99% utilization at a mere 1280 x 800 internal resolution." This doesn't mean what you think it means. The GPU shows 99% utilization because Doom is running in a single kernel launch. Any waiting between frame updates...