similar to: [PATCH] nv50/ir: allow load propagation when flags are defined

Retrieving the high 32 bits of a signed multiply is rather annoying. It appears that the simplest way to do this is to compute the absolute value of the arguments, and perform a u32 x u32 -> u64 operation. If the arguments' signs differ, then negate the result. Since there is no u64 support in the cvt instruction, we have the perform the 2's complement negation "by hand".

And you're allowing saturate/neg emission on the short form. Is this already in envytools? Also, what's the shortForm thing? This change is probably fine, but the changelog needs work. On Sat, Jan 10, 2015 at 7:22 PM, Roy Spliet <rspliet at eclipso.eu> wrote: > MAD IMM has a very specific SDST == SSRC2 requirement, so don't emit > > Signed-off-by: Roy Spliet <rspliet

Op 11-01-15 om 01:34 schreef Ilia Mirkin: > And you're allowing saturate/neg emission on the short form. Yes > Is this already in envytools? Tesla floating point instructions are poorly documented in the RST documents; fmad is no exception. I'll make sure to check envydis. > Also, what's the shortForm thing? Documented in envytools; see

Are you sure this won't break compute shaders on fermi? Could you please double-check that? One minor comment below. On 03/17/2016 05:07 PM, Hans de Goede wrote: > Some of the lowering steps we currently do for FILE_MEMORY_GLOBAL only > apply to buffers, making it impossible to use FILE_MEMORY_GLOBAL for > OpenCL global buffers. > > This commits changes the buffer code to use

On 04/08/2016 12:17 PM, Hans de Goede wrote: > Hi, > > On 23-03-16 23:10, Samuel Pitoiset wrote: >> Are you sure this won't break compute shaders on fermi? >> Could you please double-check that? > > I just checked: > > lspci: > 01:00.0 VGA compatible controller: NVIDIA Corporation GF119 [GeForce GT > 610] (rev a1) > > Before this patch-set: >

Add emission rules for negative and saturate flags for MAD 4-byte opcodes, and get rid of constraints. Short MAD has a very specific SDST == SSRC2 requirement, and since MAD IMM is short notation + 4-byte immediate, don't have the compiler create MAD IMM instructions yet. V2: Document MAD as supported short form Signed-off-by: Roy Spliet <rspliet at eclipso.eu> ---

MAD IMM has a very specific SDST == SSRC2 requirement, so don't emit Signed-off-by: Roy Spliet <rspliet at eclipso.eu> --- .../drivers/nouveau/codegen/nv50_ir_emit_nv50.cpp | 18 ++++++++++++------ .../drivers/nouveau/codegen/nv50_ir_target_nv50.cpp | 2 +- 2 files changed, 13 insertions(+), 7 deletions(-) diff --git a/src/gallium/drivers/nouveau/codegen/nv50_ir_emit_nv50.cpp

On 04/12/2016 12:04 PM, Hans de Goede wrote: > Hi, > > On 08-04-16 18:14, Samuel Pitoiset wrote: >> >> >> On 04/08/2016 12:17 PM, Hans de Goede wrote: >>> Hi, >>> >>> On 23-03-16 23:10, Samuel Pitoiset wrote: >>>> Are you sure this won't break compute shaders on fermi? >>>> Could you please double-check that?

Some of the lowering steps we currently do for FILE_MEMORY_GLOBAL only apply to buffers, making it impossible to use FILE_MEMORY_GLOBAL for OpenCL global buffers. This commits changes the buffer code to use FILE_MEMORY_BUFFER at the ir_from_tgsi and lowering steps, freeing use of FILE_MEMORY_GLOBAL for use with OpenCL global buffers. Note that after lowering buffer accesses use the

Hi, On 23-03-16 23:10, Samuel Pitoiset wrote: > Are you sure this won't break compute shaders on fermi? > Could you please double-check that? I just checked: lspci: 01:00.0 VGA compatible controller: NVIDIA Corporation GF119 [GeForce GT 610] (rev a1) Before this patch-set: [hans at plank piglit]$ ./piglit run -o shader -t '.*arb_shader_storage_buffer_object.*' results/shader

Hi, On 08-04-16 18:14, Samuel Pitoiset wrote: > > > On 04/08/2016 12:17 PM, Hans de Goede wrote: >> Hi, >> >> On 23-03-16 23:10, Samuel Pitoiset wrote: >>> Are you sure this won't break compute shaders on fermi? >>> Could you please double-check that? >> >> I just checked: >> >> lspci: >> 01:00.0 VGA compatible

Add emission rules for negative and saturate flags for MAD 4-byte opcodes, and get rid of some of the constraints. Obviously tested with a wide variety of shaders. V2: Document MAD as supported short form V3: Split up IMM from short-form modifiers Signed-off-by: Roy Spliet <rspliet at eclipso.eu> --- src/gallium/drivers/nouveau/codegen/nv50_ir_emit_nv50.cpp | 10 ++++------

FILE_MEMORY_GLOBAL is currently only used for buffer handling, as we do not yet have (opencl) global memory support. Global memory support actually requires some different handling during lowering, so rename FILE_MEMORY_GLOBAL to FILE_MEMORY_BUFFER to reflect that the current code is for buffer handling, this will allow the later (re-)addition of FILE_MEMORY_GLOBAL for regular global memory.

Hi, On 16-03-16 15:55, Ilia Mirkin wrote: > This approach leads to the emitters needing to know about both global and > buffer, even though at that point, they are identical. I was thinking that > in the lowering logic, buffer would just get rewritten as global (with the > offset added), thus not needing any change to the emitters. What do you > think about such an approach? I was

Signed-off-by: Ilia Mirkin <imirkin at alum.mit.edu> --- Untested beyond compiling a few shaders to see if they look like they might work. nvdisasm agrees with envydis's decoding of these things. Will definitely get ahold of a G200 to run tests on before pushing this. .../drivers/nouveau/codegen/nv50_ir_emit_nv50.cpp | 94 ++++++++++++++++++---

This approach leads to the emitters needing to know about both global and buffer, even though at that point, they are identical. I was thinking that in the lowering logic, buffer would just get rewritten as global (with the offset added), thus not needing any change to the emitters. What do you think about such an approach? On Mar 16, 2016 2:24 AM, "Hans de Goede" <hdegoede at

Add support for OpenCL global memory buffers, note this has only been tested with regular load and stores and likely needs more work for e.g. atomic ops. Signed-off-by: Hans de Goede <hdegoede at redhat.com> --- src/gallium/drivers/nouveau/codegen/nv50_ir.h | 1 + .../drivers/nouveau/codegen/nv50_ir_emit_gk110.cpp | 31 +++++++++++++++++-----

Hi, On 16-03-16 11:37, Samuel Pitoiset wrote: > Could you please get rid of the cosmetic changes (eg. the switch ones)? > Because this doesn't really improve readability and in my opinion these changes should be eventually done in a separate patch. I need at least halve of those cosmetic changes, because half of them is not cosmetic, e.g. : - case FILE_MEMORY_BUFFER: code[1] =

On 09.05.2015 07:35, Ilia Mirkin wrote: > This has started to happen more now that the backend is producing > KILL_IF more often. > > Signed-off-by: Ilia Mirkin <imirkin at alum.mit.edu> > --- > .../drivers/nouveau/codegen/nv50_ir_peephole.cpp | 29 ++++++++++++++++++++++ > .../nouveau/codegen/nv50_ir_target_nv50.cpp | 2 ++ > 2 files changed, 31

Could you please get rid of the cosmetic changes (eg. the switch ones)? Because this doesn't really improve readability and in my opinion these changes should be eventually done in a separate patch. Other than that, this patch is : Reviewed-by: Samuel Pitoiset <samuel.pitoiset at gmail.com> Yes, this probably won't work as is for atomic operations but the lowering pass is