Some SM69 Vectorized QuadRead and other tests fail on x86 WARP

[alsepkow]

The following tests hit OOM issues when running on WARP in x86 builds. These tests stress warps shader compiler to use too much memory for large vectors on x86 builds.

We likely just need to cap the vector size to something smaller than 1024 when running on x86 against WARP. The size 512 cases pass.
Updating the pipelines running this against x86 WARP to use the IsRITP runtime parameter is also a reasonable option.

clang-taef-exec :: ExecHLSLTests.dll/DxilConf_SM69_Vectorized_Core::AsDouble_uint32_t
clang-taef-exec :: ExecHLSLTests.dll/DxilConf_SM69_Vectorized_Core::AsUint_SplitDouble_double
clang-taef-exec :: ExecHLSLTests.dll/DxilConf_SM69_Vectorized_Core::QuadReadAcrossDiagonal_HLSLHalf_t
clang-taef-exec :: ExecHLSLTests.dll/DxilConf_SM69_Vectorized_Core::QuadReadAcrossDiagonal_float
clang-taef-exec :: ExecHLSLTests.dll/DxilConf_SM69_Vectorized_Core::QuadReadAcrossX_HLSLHalf_t
clang-taef-exec :: ExecHLSLTests.dll/DxilConf_SM69_Vectorized_Core::QuadReadAcrossX_float
clang-taef-exec :: ExecHLSLTests.dll/DxilConf_SM69_Vectorized_Core::QuadReadAcrossY_HLSLHalf_t
clang-taef-exec :: ExecHLSLTests.dll/DxilConf_SM69_Vectorized_Core::QuadReadAcrossY_float
clang-taef-exec :: ExecHLSLTests.dll/DxilConf_SM69_Vectorized_Core::QuadReadLaneAt_HLSLHalf_t
clang-taef-exec :: ExecHLSLTests.dll/DxilConf_SM69_Vectorized_Core::QuadReadLaneAt_float
clang-taef-exec :: ExecHLSLTests.dll/DxilConf_SM69_Vectorized_Core::QuadReadLaneAt_uint16_t

Failure in logs:
Compiling shader:

Target profile: cs_6_9

Arguments:  -enable-16bit-types -DTYPE=uint16_t -DNUM=1024 -DOPERATOR= -DFUNC=TestQuadRead -DFUNC_TEST_QUAD_READ=1 -DNUMTHREADS_XYZ=2,2,1 -DOP_STORES_RESULT_ON_SPECIFIC_LANE=1 -DQUAD_READ_FUNC=QuadReadLaneAt -DSOURCE_LANE_ID=2 -DOUT_TYPE=uint16_t -DBASIC_OP_TYPE=0x2

Device lost: 0x887a0005 (DXGI_ERROR_DEVICE_REMOVED)