Re: [Mesa-dev] [PATCH] R600/SI: Split global vector loads with more than 4 elements

2014-02-11 Thread Tom Stellard
On Mon, Feb 10, 2014 at 02:35:03PM -0800, Matt Arsenault wrote: > Why would you want to do this for the small types? You should be able to > load those in fewer loads and then promote them. > We only custom lower v*i32 loads, so this code won't be executed on smaller types. -Tom > On 02/10/2014

Re: [Mesa-dev] [PATCH] R600/SI: Split global vector loads with more than 4 elements

2014-02-10 Thread Aaron Watry
Hi Tom, This definitely fixes some issues that I've been seeing with int8/16 vload8() and vload16() in CL. vstore8/vstore16 are still broken, but at least the loads are working now (I've only tested int, but I can give a full test run if you want/need). For reference, the tests that failed befor

Re: [Mesa-dev] [PATCH] R600/SI: Split global vector loads with more than 4 elements

2014-02-10 Thread Matt Arsenault
Why would you want to do this for the small types? You should be able to load those in fewer loads and then promote them. On 02/10/2014 01:32 PM, Tom Stellard wrote: From: Tom Stellard --- lib/Target/R600/SIISelLowering.cpp | 8 +- test/CodeGen/R600/load.ll | 178 +++

[Mesa-dev] [PATCH] R600/SI: Split global vector loads with more than 4 elements

2014-02-10 Thread Tom Stellard
From: Tom Stellard --- lib/Target/R600/SIISelLowering.cpp | 8 +- test/CodeGen/R600/load.ll | 178 +++-- 2 files changed, 98 insertions(+), 88 deletions(-) diff --git a/lib/Target/R600/SIISelLowering.cpp b/lib/Target/R600/SIISelLowering.cpp index 9537