TE
TechEcho
Home24h TopNewestBestAskShowJobs
GitHubTwitter
Home

TechEcho

A tech news platform built with Next.js, providing global tech news and discussions.

GitHubTwitter

Home

HomeNewestBestAskShowJobs

Resources

HackerNews APIOriginal HackerNewsNext.js

© 2025 TechEcho. All rights reserved.

Intel Xe-LP GPU Architecture Deep Dive

50 pointsby m3atalmost 5 years ago

2 comments

pjmlpalmost 5 years ago
&gt; The downside is that it also means that Intel is the only hardware vendor launching a new GPU&#x2F;architecture in 2020 without support for the next generation of features, which Microsoft &amp; co are codifying as DirectX 12 Ultimate. The consumer-facing trade name for feature level 12_2, DirectX Ultimate incorporates support for variable rate shading tier 2, along with ray tracing, mesh shaders, and sampler feedback. And to be fair to Intel, expecting ray tracing in an integrated part in 2020 was always a bit too much of an ask. But some additional progress would always be nice to see. Plus it puts DG1 in a bit of an odd spot, since it’s a discrete GPU without 12_2 functionality.<p>Not really on a good starting foot.
评论 #24171086 未加载
评论 #24171645 未加载
评论 #24171283 未加载
dragontameralmost 5 years ago
CPU cores are pretty much a done deal IMO (unless Nuvia&#x27;s hype is real. I guess we&#x27;ll see about that soon). For the last 10, maybe 20 years, CPUs have simply grown bigger and have more or less done the same features.<p>CPUs simply push wider SIMD (SSE -&gt; AVX -&gt; AVX512), bigger out-of-execution buffers, larger registers, deeper branch prediction, more cache. The biggest advancement seemed to be Intel&#x27;s uop-cache in Sandy Bridge (a concept that was first tried in the Pentium4, so not really a &quot;new&quot; innovation).<p>Even ARM chips are more of the same. Wider SIMD (see A64fx: #1 supercomputer in the world for 512-bit SVE), deeper out-of-order execution, better branch prediction. Its rare to see something new<p>----------<p>In contrast, whenever I see GPU-architectures, everything is so hugely different. True, GPUs all have a &quot;SIMD focus&quot;, but there&#x27;s so many exciting things about SIMD that just haven&#x27;t been explored yet.<p>NVidia has Tensor-cores, Raytracing cores. It was already crazy that NVidia had a full matrix-multiplication implemented in its FP16 processor in Volta, but Ampere is now pushing sparse matrix multiplication at the chip level (!!).<p>AMD GCN and RDNA push GPU&#x2F;SIMD architecture in a different direction. While a chunk of RDNA can be described as &quot;more of the same&quot;, the &quot;SubVector execution&quot; feature grossly changes how Wave64 wavefronts are executed to save on register space and potentially have a better memory access pattern. Innovation continues in the AMD space for sure.<p>Intel finally enters with their GPU &#x2F; SIMD architecture, and they start of with... SIMD8 + SIMD2 (instead of the Wave32 of RDNA&#x2F;NVidia or Wave64 for GCN). Finally, Intel is minimizing hardware scheduling (saving on die area), and focusing on software to schedule the instructions.<p>Three very, very different architectures. Three very different ideologies for how SIMD compute &#x2F; GPUs should be built. The SIMD &#x2F; GPU architecture space remains more innovative than the CPU-architecture space by a longshot. Its exciting to see.<p>--------<p>I&#x27;m not saying that CPU-advancements are dead per se. AMD did make chiplets a hot commodity (but chiplets were already experimented with Power5, and chiplets seem like a natural evolution to dual-socket or NUMA architectures... just cheap enough for a consumer now). Intel&#x27;s new Tremont Atom core has this cool &quot;dual-decoder&quot; design (where one decoder works on one-branch, and the 2nd decoder works on a 2nd branch in parallel). There are certainly new ideas being tested in the CPU world. But nothing as majorly game changing as what Intel &#x2F; AMD &#x2F; NVidia are doing to the GPU space. (Or Google if we include TPUs in the discussion. A TPU isn&#x27;t SIMD, but it shares some degree of similarity, especially to NVidia&#x27;s Tensor cores)<p>I guess big.LITTLE (and Intel&#x27;s Lakefield platform) was a big change in the last 10 years. But that hasn&#x27;t really seen much success in the Desktop&#x2F;Server markets.
评论 #24173866 未加载
评论 #24172303 未加载