Final opcodes by tlively · Pull Request #452 · WebAssembly/simd

tlively · 2021-02-06T02:00:04Z

A draft numbering of the most recently added instructions. I tried to keep related instructions together, so there are still holes in the opcode space. I also could have more aggressively filled holes at the expense of destroying much of the internal structure of the opcode space. There are 236 instructions, so in principle we could fit them all below 0x100, but it would be ugly. Comments welcome :)

lars-t-hansen · 2021-02-08T09:15:54Z

proposals/simd/NewOpcodes.md

 | i8x16.sub_sat_s      | 0x72   | i16x8.sub_sat_s          | 0x92   | ---- sub_sat ----        | 0xb2   | -------------            | 0xd2   |
 | i8x16.sub_sat_u      | 0x73   | i16x8.sub_sat_u          | 0x93   | ---- sub_sat ----        | 0xb3   | -------------            | 0xd3   |
-| -------------        | 0x74   | -------------            | 0x94   | -------------            | 0xb4   | -------------            | 0xd4   |
+|                      | 0x74   | -------------            | 0x94   | -------------            | 0xb4   | -------------            | 0xd4   |


Spurious change from dashes to blank?

Indeed, fixed!

Maratyszcza · 2021-02-08T17:51:29Z

Why are 0xe2 and 0xee reserved for round?

Maratyszcza · 2021-02-08T17:54:52Z

IMO we should recycle at least the opcodes where the original reserved instruction is meaningless. E.g. 0x67-0x6a opcodes are meaningless for widen instructions. Also, AFAICT it is missing v128.any_true.

proposals/simd/NewOpcodes.md

tlively · 2021-02-08T21:00:17Z

@Maratyszcza I moved v128.any_true to 0x53 because it's now a one-off instruction, but if that renumbering would be disruptive, I could move it back to where it was.

yurydelendik · 2021-02-08T22:44:11Z

proposals/simd/BinarySIMD.md

+| `i64x2.le_s`                    |   `0x106`| -                        |
+| `i64x2.ge_s`                    |   `0x107`| -                        |
+| `i64x2.all_true`                |    `0xc3`| -                        |
+| `f64x2.convert_low_i32x4_s`     |    `0xfc`| -                        |


fc and fd are taken by v128.load32_zero and v128.load64_zero

Thanks for catching that!

tlively · 2021-02-09T00:58:55Z

Ok, I pushed a new version based on @ngzhian's suggestion and compacting more instructions into previous opcode holes. This version only uses opcodes up to 0xff, which will simplify V8 at least. The only block of instructions that really got scattered were the f64x2 rounding instructions, which are scattered among the integer operations.

proposals/simd/NewOpcodes.md

ngzhian · 2021-02-09T01:06:55Z

Amazing bin packing work Thomas, thanks!

yurydelendik · 2021-02-09T14:59:12Z

proposals/simd/BinarySIMD.md

+| `i64x2.extmul_high_i32x4_s`     |    `0xdd`| -                        |
+| `i64x2.extmul_low_i32x4_u`      |    `0xde`| -                        |
+| `i64x2.extmul_high_i32x4_u`     |    `0xdf`| -                        |
+| `i16x8.q15mulr_sat_s`           |    `0x9c`| -                        |


I believe it is 0x82 (i16x8.extmul_low_i8x16_s uses 0x9c)

Thanks, fixed.

tlively · 2021-02-09T21:32:16Z

Why are 0xe2 and 0xee reserved for round?

TBH I don't remember. I must have thought at the time of the last renumbering that there would be just one floating point rounding instruction for each of f32x4 and f64x2.

zeux · 2021-02-11T16:18:42Z

256 opcodes was indeed enough for anyone 🚀

proposals/simd/NewOpcodes.md

tlively · 2021-02-19T18:14:18Z

Merging because there were no further comments on this at today's meeting.

Opcodes were finalized in WebAssembly#452, this updates the interpreter and spec text to use these final opcodes.

See WebAssembly/simd#452 for finalized opcodes and WebAssembly/simd#467 for some renamings.

Opcodes were finalized in #452, this updates the interpreter and spec text to use these final opcodes.

See WebAssembly/simd#452 for finalized opcodes and WebAssembly/simd#467 for some renamings.

tlively added 2 commits February 5, 2021 17:55

Final opcodes

45112d3

Add ops from #232 to NewOpcodes.md

7633dd0

lars-t-hansen reviewed Feb 8, 2021

View reviewed changes

yurydelendik mentioned this pull request Feb 8, 2021

New (and renumbered) opcodes wasdk/wasmparser#90

Closed

Undo spurious change

37de1c3

ngzhian reviewed Feb 8, 2021

View reviewed changes

proposals/simd/NewOpcodes.md Outdated Show resolved Hide resolved

yurydelendik reviewed Feb 8, 2021

View reviewed changes

Address feedback, compact to 0xff

2cabfd4

ngzhian reviewed Feb 9, 2021

View reviewed changes

proposals/simd/NewOpcodes.md Show resolved Hide resolved

ngzhian reviewed Feb 9, 2021

View reviewed changes

proposals/simd/NewOpcodes.md Outdated Show resolved Hide resolved

Swap truc_sat_zero and convert_low

01b5f01

yurydelendik reviewed Feb 9, 2021

View reviewed changes

yurydelendik mentioned this pull request Feb 9, 2021

Final SIMD opcodes bytecodealliance/wasm-tools#222

Merged

1 task

ngzhian mentioned this pull request Feb 9, 2021

Fix ext mul opcodes in BinarySIMD.md #406

Closed

Fix q15 mul opcode

bbb3b5f

tlively mentioned this pull request Feb 16, 2021

Agenda for sync meeting 2/19/21 #470

Closed

ngzhian approved these changes Feb 16, 2021

View reviewed changes

proposals/simd/NewOpcodes.md Show resolved Hide resolved

tlively merged commit 2c28fa8 into master Feb 19, 2021

alexcrichton deleted the final-opcodes branch February 19, 2021 18:45

alexcrichton restored the final-opcodes branch February 19, 2021 18:45

ngzhian added a commit to ngzhian/simd that referenced this pull request Feb 26, 2021

Update interpreter and text with finalized opcodes

d303f63

Opcodes were finalized in WebAssembly#452, this updates the interpreter and spec text to use these final opcodes.

ngzhian mentioned this pull request Feb 26, 2021

Update interpreter and text with finalized opcodes #486

Merged

syrusakbary mentioned this pull request Mar 2, 2021

Final SIMD Proposal wasmerio/wasmer#2156

Closed

ngzhian added a commit to ngzhian/wasmparser that referenced this pull request Mar 2, 2021

fix(simd): update SIMD opcodes

737b206

See WebAssembly/simd#452 for finalized opcodes and WebAssembly/simd#467 for some renamings.

ngzhian mentioned this pull request Mar 2, 2021

fix(simd): update SIMD opcodes wasdk/wasmparser#92

Merged

ngzhian added a commit to ngzhian/wasmparser that referenced this pull request Mar 2, 2021

fix(simd): update SIMD opcodes

3678aa1

See WebAssembly/simd#452 for finalized opcodes and WebAssembly/simd#467 for some renamings.

ngzhian added a commit to ngzhian/wasmparser that referenced this pull request Mar 3, 2021

fix(simd): update SIMD opcodes

ddca519

See WebAssembly/simd#452 for finalized opcodes and WebAssembly/simd#467 for some renamings.

ngzhian added a commit that referenced this pull request Mar 3, 2021

Update interpreter and text with finalized opcodes

e5ce801

Opcodes were finalized in #452, this updates the interpreter and spec text to use these final opcodes.

bmeurer pushed a commit to wasdk/wasmparser that referenced this pull request Mar 4, 2021

fix(simd): update SIMD opcodes

648fff5

See WebAssembly/simd#452 for finalized opcodes and WebAssembly/simd#467 for some renamings.

ngzhian mentioned this pull request Mar 10, 2021

Tracking instructions with unassigned opcodes #421

Closed

Conversation

tlively commented Feb 6, 2021

Uh oh!

lars-t-hansen Feb 8, 2021

Choose a reason for hiding this comment

Uh oh!

tlively Feb 8, 2021

Choose a reason for hiding this comment

Uh oh!

Maratyszcza commented Feb 8, 2021

Uh oh!

Maratyszcza commented Feb 8, 2021 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

tlively commented Feb 8, 2021

Uh oh!

yurydelendik Feb 8, 2021

Choose a reason for hiding this comment

Uh oh!

tlively Feb 8, 2021

Choose a reason for hiding this comment

Uh oh!

tlively commented Feb 9, 2021

Uh oh!

Uh oh!

Uh oh!

ngzhian commented Feb 9, 2021

Uh oh!

yurydelendik Feb 9, 2021

Choose a reason for hiding this comment

Uh oh!

tlively Feb 9, 2021

Choose a reason for hiding this comment

Uh oh!

tlively commented Feb 9, 2021

Uh oh!

zeux commented Feb 11, 2021 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

tlively commented Feb 19, 2021

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

6 participants

Maratyszcza commented Feb 8, 2021 •

edited

Loading

zeux commented Feb 11, 2021 •

edited

Loading