https://sourceware.org/bugzilla/show_bug.cgi?id=32813
--- Comment #4 from Andreas Abel ---
With respect to performance, the processor behaviors are not identical. The one
with an 64-bit operand size is one byte longer, which can have various effects
on performance.
Yes, I could add workarounds t
https://sourceware.org/bugzilla/show_bug.cgi?id=32813
--- Comment #5 from Andreas Abel ---
If it is not a bug, but a conscious choice to prefer the form without the REX
prefix to optimize the instruction size, I would expect that `{nooptimize} LSL
RCX, RDX` has the prefix, but this is not the cas
https://sourceware.org/bugzilla/show_bug.cgi?id=32813
--- Comment #2 from Andreas Abel ---
Yes, it causes issues in
https://github.com/andreas-abel/nanoBench/tree/master/tools/cpuBench, which is
the code that generates the benchmarks that are behind https://www.uops.info/
Because of the first is
https://sourceware.org/bugzilla/show_bug.cgi?id=32813
--- Comment #8 from Andreas Abel ---
I do not need to support other assemblers; this was just to show how that the
current version of gas is the outlier here. And I would not need workarounds
for the other assemblers, as they encode the instru
Assignee: unassigned at sourceware dot org
Reporter: abel at cs dot uni-saarland.de
Target Milestone: ---
Example:
```
.intel_syntax noprefix
LSL ECX, EDX
LSL RCX, RDX
REX.W LSL ECX, EDX
```
objdump output:
```
0: 0f 03 calslecx,edx
3: 0f 03 ca
https://sourceware.org/bugzilla/show_bug.cgi?id=24076
Andreas Abel changed:
What|Removed |Added
CC||abel at cs dot uni-saarland.de
Component: gas
Assignee: unassigned at sourceware dot org
Reporter: abel at cs dot uni-saarland.de
Target Milestone: ---
For the instruction "lwpins rax,DWORD PTR [rax],0x1", gas reports "Error: no
instruction mnemonic suffix given; can't determine immediate size&
