1738744404414759381

2026-08-26 18:50:44 -04:00
1 parent 11e36d4d99
commit a0b2a3a8f7
@@ -0,0 +1,37 @@
+---
+title: "@NOTimothyLottes Do you know other equivalent graphs or ressources for other GPUs (GTX, RTX, Snapdragon XR, ...) ?"
+type: archive
+source: twitter
+source_url: "https://x.com/BrettRidel/status/1738843423891755098"
+author: "Brett Ridel"
+handle: BrettRidel
+post_id: "1738843423891755098"
+date: 2023-12-24
+archived: 2026-08-26
+draft: false
+tags:
+  - archive
+  - twitter
+  - NOTimothyLottes
+description: "@NOTimothyLottes Do you know other equivalent graphs or ressources for other GPUs (GTX, RTX, Snapdragon XR, ...) ?"
+in_reply_to: ""
+parent_post_id: "1738744404414759381"
+---
+
+## Source
+
+- URL: https://x.com/BrettRidel/status/1738843423891755098
+- Author: Brett Ridel (@BrettRidel)
+- Posted: 2023-12-24 08:45:53
+
+## Branch
+
+**1/** **@BrettRidel** ^1738843423891755098
+
+**@NOTimothyLottes**
+
+Do you know other equivalent graphs or ressources for other GPUs (GTX, RTX, Snapdragon XR, ...) ?
+
+## Related
+
+- Spine: [[archive/threads/NOTimothyLottes/2023-12-24-two-great-graphs-from-https-chipsandcheese-com]]
@@ -0,0 +1,53 @@
+---
+title: "Two great graphs from https://chipsandcheese.com/2023/12/23/nintendo-switchs-igpu-maxwell-nerfed-edition/ - Why I always manually vectorize 16-bit code, and also why you always need to manually check disassembly."
+type: archive
+source: twitter
+source_url: "https://x.com/NOTimothyLottes/status/1738744404414759381"
+author: "NOTimothyLottes"
+handle: NOTimothyLottes
+post_id: "1738744404414759381"
+date: 2023-12-24
+archived: 2026-08-26
+draft: false
+tags:
+  - archive
+  - twitter
+  - NOTimothyLottes
+description: "Two great graphs from https://chipsandcheese.com/2023/12/23/nintendo-switchs-igpu-maxwell-nerfed-edition/ - Why I always manually vectorize 16-bit code, and also why you always need to manually check disassembly."
+in_reply_to: ""
+---
+
+## Source
+
+- URL: https://x.com/NOTimothyLottes/status/1738744404414759381
+- Author: NOTimothyLottes (@NOTimothyLottes)
+- Posted: 2023-12-24 02:12:24
+
+## Thread
+
+**1/** **@NOTimothyLottes** ^1738744404414759381
+
+Two great graphs from https://chipsandcheese.com/2023/12/23/nintendo-switchs-igpu-maxwell-nerfed-edition/ - Why I always manually vectorize 16-bit code, and also why you always need to manually check disassembly.
+
+![](https://pbs.twimg.com/media/GCFBbyUWIAAaWoR?format=png&name=orig)
+![](https://pbs.twimg.com/media/GCFBdy9WEAE0JYk?format=png&name=orig)
+
+Branches: [[archive/threads/NOTimothyLottes/2023-12-24-two-great-graphs-from-https-chipsandcheese-com/2023-12-24-BrettRidel-do-you-know-other-equivalent-graphs-or-ressources]]
+
+**2/** **@AgileJebrim** ^1738750536302670083
+
+**@NOTimothyLottes**
+
+Why are float16s so much faster than int16s?
+
+**3/** **@NOTimothyLottes** ^1738756633738522752
+
+**@AgileJebrim**
+
+Same bit width, FMA needs smaller multipler than IMAD. Thus HW mul24 fast paths that never get exposed. But 16-bit is small enough already. I thought on AMD, packed 16-bit int or float are the same perf. Some vendors though try area savings by de-throughputing integer/etc ALUs
+
+**4/** **@AgileJebrim** ^1738757938376761652
+
+**@NOTimothyLottes**
+
+Oh are you only comparing FMA? If it was just additions, would they be on par?
@@ -37,6 +37,7 @@ tags:
 - [[archive/threads/NOTimothyLottes/2023-12-20-ideally-a-gpu-profiling-graph-would-provide-at]]
 - [[archive/threads/NOTimothyLottes/2023-12-21-code-shy-brain-says-its-strange-to-be-reading]]
 - [[archive/threads/NOTimothyLottes/2023-12-24-the-sometimes-extra-2-5-seconds-on-vulkan]]
+- [[archive/threads/NOTimothyLottes/2023-12-24-two-great-graphs-from-https-chipsandcheese-com]]
 - [[archive/threads/NOTimothyLottes/2023-12-25-virtualized-xmas-thread]]
 - [[archive/threads/NOTimothyLottes/2023-12-26-no-luck-getting-d3dkmtgetscanline-to-work-thus-far]]
 - [[archive/threads/NOTimothyLottes/2023-12-29-end-of-2023-scaling-parallel-gfx-and-memory-thread]]