issues
search
WebAssembly
/
relaxed-simd
Relax the strict determinism requirements of SIMD operations.
Other
38
stars
6
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
Relation to NaN non-determinism
#154
RalfJung
opened
8 months ago
11
Seeming contradiction in summary regarding the nature of non-determinism in this proposal
#153
RalfJung
opened
8 months ago
0
What is the intended workflow of actually leveraging any implementation defined instructions?
#152
juj
opened
9 months ago
5
Fix relaxed trunc pseudocode in Overview
#151
thibaudmichaud
closed
9 months ago
1
Does deterministic profile have threads instructions?
#150
eqrion
opened
9 months ago
29
Document that "either" comes from threads proposal
#149
ngzhian
closed
10 months ago
2
1. What are the instructions being proposed?
#148
HxYxI
closed
1 year ago
0
URL
#147
HxYxI
closed
1 year ago
0
Esim
#146
HxYxI
closed
1 year ago
2
Fix relaxed laneselect to allow looking at top bit of byte
#145
ngzhian
closed
10 months ago
3
Add more allowed return values for relaxed trunc unsigned
#144
ngzhian
closed
1 year ago
4
SIMD subgroup meeting on 2023-06-02
#143
ngzhian
closed
1 year ago
2
Next SIMD sync
#142
penzn
closed
1 year ago
2
Behavior of relaxed operations for multi-tier compilers
#141
yurydelendik
opened
1 year ago
1
Alternative output for i32x4_relaxed_trunc.wast tests
#140
yurydelendik
opened
1 year ago
7
Merge from upstream spec
#139
ngzhian
closed
1 year ago
0
SIMD subgroup meeting on 2023-04-14
#138
ngzhian
closed
1 year ago
4
SIMD subgroup meeting on 2023-03-17
#137
ngzhian
closed
1 year ago
6
Add prose for exec numerics and fma
#135
ngzhian
closed
1 year ago
2
Phase 4 tracking issue
#134
dtig
opened
1 year ago
4
Update overview spec of `laneselect` instructions
#133
alexcrichton
closed
1 year ago
0
Fix typo in relaxed dot test
#132
ngzhian
closed
1 year ago
0
Fix some tests in `i32x4_relaxed_trunc.wast`
#131
alexcrichton
closed
1 year ago
2
Update int dot tests to allow for saturation of intermediate 16-bit
#130
ngzhian
closed
1 year ago
1
Rounding semantics for `i32x4.relaxed_dot_i8x16_i7x16_add_s`
#129
alexcrichton
closed
1 year ago
5
Merge from upstream spec
#128
ngzhian
closed
1 year ago
0
Fix relaxed trunc signed in overview
#127
ngzhian
closed
1 year ago
0
Float-to-signed-integer conversion in Overview.md vs spec
#126
alexcrichton
closed
1 year ago
2
Codegen of i16x8.relaxed_laneselect
#125
alexcrichton
opened
1 year ago
28
Fix relaxed nmadd tests
#124
ngzhian
closed
1 year ago
0
Add expected results for relaxed dot
#123
ngzhian
closed
1 year ago
0
Check multiple calls to relaxed instructions return same results
#122
ngzhian
closed
1 year ago
0
Fix expected result type in relaxed dot product tests
#121
ngzhian
closed
1 year ago
0
Change relaxed madd in interpreter to be FMA
#120
ngzhian
closed
1 year ago
0
Fix names of dot instructions
#119
ngzhian
closed
1 year ago
0
Remove bf16 dot
#118
ngzhian
closed
1 year ago
0
Discrepancy between overview's dot product instructions and tests
#117
alexcrichton
closed
1 year ago
2
SIMD subgroup meeting on 2023-02-17
#116
ngzhian
closed
1 year ago
1
Semantics on top of profiles
#115
ngzhian
closed
1 year ago
6
SIMD subgroup meeting on 2023-01-20
#114
ngzhian
closed
1 year ago
3
Use profiles for updates to spec for relaxed mode
#113
ngzhian
closed
1 year ago
1
Merge upstream
#112
ngzhian
closed
1 year ago
0
Remove bf16 from overview and implementation status
#111
ngzhian
closed
1 year ago
0
Add tests to ensure that multiple calls to the same instruction return same results
#110
ngzhian
closed
1 year ago
0
SIMD subgroup meeting on 2022-11-18
#109
ngzhian
closed
1 year ago
2
Deterministic `NaN`
#108
penzn
opened
1 year ago
8
Feedback and next steps on relaxed mode
#107
ngzhian
opened
1 year ago
18
Poll on relaxed mode bf16 dot product
#106
ngzhian
closed
1 year ago
9
Poll on relaxed mode
#105
ngzhian
opened
1 year ago
6
Floating point vs integer and fixed point
#104
penzn
opened
1 year ago
2
Next