AMD Radeon Vega 64

And my 6-year-old GPU (Vega64) has 4096 32-bit multipliers at 1500 MHz clock. (And I bought it a year or two late, so it was only $400). With 8GB HBM2 RAM at 512GBps (yes, Gigabytes per second) throughput. You absolutely are not getting anything close to a consumer GPU in terms of performance or cost on any other platform. --------- FPGAs advantage is that systolic array arrangement. I'm pretty sure GPUs will win in a "hardware multiplier" war, but it might be harder to get all the data into a typical GPU There's a fair number of programs that will never utilize a GPU with any level of efficiency. That's where FPGAs come in, utilization will be better because of LUTs and Routers and all that magic. But if people are looking for a TFLOP fight with raw numbers of multipliers? My bet is on the GPU and SIMD-processing in general. Those architectures are just crazy. -------------- BTW: That top end DSP slice is documented here: https://docs.xilinx.com/v/u/en-US/ug579-ultrascale-dsp That's a 27-bit x 18-bit multiplier per dsp. In contrast, a 32-bit floating point operation is 24-bit x 24-bit multiplier plus a bit of extra stuff for the exponent bits. I'm pretty sure that the 4096-shader Vega64 actually had 32-bit multipliers but I'm not 100% sure on that. But the float 24-bit x 24-bit multiplier would need more than one dsp to replicate, maybe two of them. AMD's newest GPUs are 6144-shader... but those shaders can execute _TWO_ multiplies per clock tick for a total of 12288 hardw…

I'm not usually one to whine, but agreed; additionally, add contrast to the modifiers (e.g. processor select). First thing I did when I visited was scale the website to 150% Super impressive comparisons, and correlates with my perception having three seperate generations of GPU (from your list pulldown). Thanks for including the "old AMD" Polaris chipsets, which are actually still much faster than lower-spec Apple silicon. I have Ollama3.1 on a VEGA64 and it really is twice as fast as an M2Pro ... ---- For anybody that thinks installing a local LLM is complicated: it's not (so long as you have more than one computer, don't tinker on your primary workhorse). I am a blue collar electrician (admittedly: geeky); no more difficult than installing linux.

On fedora (38), there is rocm-hip package. It is relatively new one, it appeared this june. For blender you will also need rocm-hip-devel, since blender dlopens() libamdhip64 not by soname (libamdhip64.so.5), but by the name libamdhip64.so, and that symlink is in the devel package. For what I can say, it does work with Vega64 and cycles is rougly twice as fast compared to cpu (ryzen 2920x in my case).

> Shuffle is the SIMD's killer app A shame that AVX512 only has pshufb (aka: permute), and is missing the GPU-instruction "bpermute", aka backwards permute. pshufb is effectively a "gather" instruction over a AVX register. Equivalent to GPU permutes. bpermute, in GPU land, is a "scatter" instruction over a vector register. There's no CPU / AVX equivalent of it. But I keep coming up with good uses of the bpermute instruction (much like pshufb is crazy flexible, its inverse, the backwards permute, is also crazy flexible). -------- Almost any code that's finding itself "gathering" data across a vector register, will inevitably "scatter" the data back at some point. Much like how "pext" is the "gather" instruction for 64-bits, you need pdep to handle the equal-and-opposite case. Its incredibly silly that AVX / AVX512 has implemented only one-half of this concept (gather / pshufb / aka Permute). I wish for the day that Intel/AMD implements (scatter / backwards-pshufb / aka Backwards-Permute). ------- Fortunately, I got Vega64 and NVidia Graphics Cards with both permute and bpermute instructions for high-speed shuffling of data. But CPU-space should benefit from this concept too.

GPUs even faster. Vega64 can explore the entire 32-bit space roughly 1-thousand times per second. (4096 shaders, each handling a 32-bit number per clock tick, albeit requiring 16384 threads to actually utilize all those shaders due to how the hardware works, at 1200 MHz) One of my toy programs was brute forcing all 32-bit constants looking for the maximum amount of "bit-avalanche" in my home-brew random number generators. It only takes a couple of seconds to run on the GPU, despite exhaustive searching and calculating the RNG across every possible 32-bit number and running statistics on the results.

> nvidia 3080... vast majority of hardware out there Uhm, no. Nvidia GPUs are not the vast majority. Unless you are gamer and talking about other gamers, then it might be right, but otherwise, Intel is the wast majority and Intel has had Wayland working fine since forever now. Intel is not alone in this, either. My AMD Vega64 was running Wayland fine since I bought it in 2017. So if your Nvidia has a problem, ask the people you paid your money to, when they want to solve it.

I only run Windows inside a VM now. The main reason I didn't switch to Linux sooner was because of games, but GPU passthrough works well enough now that I can dedicate my main GPU to a VM and get close to native performance (within 3-5%). Currently I have a Vega 64 configured for passthrough on a Threadripper 1950X system and it runs great. It's much more convenient than dual booting.