Commit f956e7f
authored
Summary:
The old `s_memtime` instruction was supported until the GFX10
architecture. Although this instruction has a higher latency than the
new shader counter, it's much more usable as a processor clock as it is
a full 64-bit counter. The new shader counter is only a 20-bit counter,
which makes it difficult to use as a standard cycle counter as it will
overflow in a few milliseconds. This patch suggests preferring
`s_memtime` for this instrinsic if it is still available.
1 parent d2565bb commit f956e7f
File tree
2 files changed
+2
-4
lines changed- llvm
- lib/Target/AMDGPU
- test/CodeGen/AMDGPU
2 files changed
+2
-4
lines changed| Original file line number | Diff line number | Diff line change | |
|---|---|---|---|
| |||
1065 | 1065 | | |
1066 | 1066 | | |
1067 | 1067 | | |
1068 | | - | |
1069 | | - | |
1070 | 1068 | | |
1071 | 1069 | | |
1072 | 1070 | | |
| |||
| Original file line number | Diff line number | Diff line change | |
|---|---|---|---|
| |||
4 | 4 | | |
5 | 5 | | |
6 | 6 | | |
7 | | - | |
8 | | - | |
| 7 | + | |
| 8 | + | |
9 | 9 | | |
10 | 10 | | |
11 | 11 | | |
| |||
0 commit comments