Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Skinny Tensors #8

Closed
guacamoleo opened this issue Apr 22, 2016 · 1 comment
Closed

Skinny Tensors #8

guacamoleo opened this issue Apr 22, 2016 · 1 comment

Comments

@guacamoleo
Copy link
Contributor

Solution Selection Logic doesn't yet check if a problem is skinny and choose a skinny solution for it.
Tests exist from Google and others for skinnyness.
GEMV is probably also a skinny tensor contraction.

@guacamoleo
Copy link
Contributor Author

done.

sdquiring added a commit to sdquiring/Tensile that referenced this issue Feb 8, 2021
# This is the 1st commit message:

Add support for GFX1030.

 - Add arch to supported list and enumerations
 - Ignore MFMA tests on gfx1030
 - Create MAC_F32 component
 - Add MACInstruction kernel property to distinguish between MAC and FMA
 - Add wave64 flag for source kernel compilation

# The commit message #2 will be skipped:

# Getting rid of unused per-arch memory instructions

# The commit message #3 will be skipped:

# Adding wave32 support.

# The commit message ROCm#4 will be skipped:

# Wait for loads if LoopIters == 1

# The commit message ROCm#5 will be skipped:

# Fixing asm-new.sh generation with wave32.

# The commit message ROCm#6 will be skipped:

# Include new enumeration values in serialization.

# The commit message ROCm#7 will be skipped:

# Skip new MFMA tests on gfx10.

# The commit message ROCm#8 will be skipped:

# Fixing .gitignore

# The commit message ROCm#9 will be skipped:

# Use default wave size.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

1 participant