[Mesa-dev] [PATCH 0/8] llvmpipe: Fixes for PowerPC/big-endian machines

[Adhemerval Zanella]

On 11/27/2012 04:28 PM, Jose Fonseca wrote:
>
> ----- Original Message -----
>> Hi all,
>>
>> This set of patches fixes all the llvmpipe lp_test_* for PowerPC. The
>> first five
>> add Altivec intrinsics that fixes all arith testcases, while the
>> remaining one
>> deals mainly with big-endian memory loads and shift/mask algorithms.
>>
>> The first three are repost of previous patches that I included just
>> for organization.
>>
>> [PATCH 1/8] PowerPC: Altivec pack/unpack intrisics
>> [PATCH 2/8] PowerPC: Altivec vector max/min intrisics
>> [PATCH 3/8] PowerPC: Altivec vector add/sub intrisics
>> [PATCH 4/8] PowerPC: Altivec floating-point rounding
>> [PATCH 5/8] PowerPC: clear Altivec NJ bit
> D3D10 spec ( http://msdn.microsoft.com/en-us/library/windows/desktop/cc308050.aspx ) states that "Denorms are flushed to sign-preserved zero on input and output of any floating-point mathematical operation.".
>
> So we might actually need to go back on this.

I added this denormal handling to fix two issues I noted in PPC tests:

1. lp_test_arit:
fract(1.40129846e-45): ref = 1.40129846e-45, out = 0, precision = -0.000000 bits, FAIL
fract(-1.40129846e-45): ref = 0.99999994, out = 0, precision = -0.000000 bits, FAIL
fract(5.8799997e-39): ref = 5.8799997e-39, out = 0, precision = -0.000000 bits, FAIL

2. lp_test_format:
Testing PIPE_FORMAT_R16_FLOAT (float) ...
FAILED
  Packed: ff 03 00 00
  Unpacked (0,0): 0 0 0 1 obtained
                  6.09756e-05 0 0 1 expected

First one is in fact related on how the Altivec fused multiplication-add is being
generated by LLVM (the addedum is not -0.0 to preseve signal, but just 0.0) and I'll
correct it in the backend.

The second issue is how the denormals are handled and the problem is with
'lp_build_half_to_float' function: it scales the exponent with a multiplication
(line 194 - LLVMBuildFMul(builder, shifted, f32_magic, "") and if the argument is
denormal it will be flushed to 0 (which it is the case of the failure). I didn't
dig into how it is handled in X86 backend,
but it does not fail in this platform.


>
>> [PATCH 6/8] llvmpipe: Fix vector constant for shuffle
>> [PATCH 7/8] llvmpipe: Add byte-swap construct calls
>> [PATCH 8/8] llvmpipe: Fix format manipulation for big-endian
> Otherwise the series looks good AFAICT. It's nice seing llvmpipe to support other ISAs!
>
> So you gave commit access, or should I commit for you?

I don't have commit access, so I'd be grateful if you commit.

>
>
> Jose
>