Daniel/master-thesis
1
0
Fork 0
Code Issues 3 Pull Requests Actions Packages Projects 1 Releases Wiki Activity

benchmarking: tuned interpreter blocksize
Some checks are pending
CI / Julia ${{ matrix.version }} - ${{ matrix.os }} - ${{ matrix.arch }} - ${{ github.event_name }} (x64, ubuntu-latest, pre) (push) Waiting to run
Details
CI / Julia ${{ matrix.version }} - ${{ matrix.os }} - ${{ matrix.arch }} - ${{ github.event_name }} (x64, ubuntu-latest, 1.10) (push) Waiting to run
Details
CI / Julia ${{ matrix.version }} - ${{ matrix.os }} - ${{ matrix.arch }} - ${{ github.event_name }} (x64, ubuntu-latest, 1.6) (push) Waiting to run
Details

Browse Source
This commit is contained in:
Daniel
2025-05-20 09:05:35 +02:00
parent a9ffd5da63
commit 250deb334c
5 changed files with 26 additions and 15 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
package/src/Interpreter.jl
View File

@ -25,7 +25,7 @@ function interpret(cudaExprs, numExprs::Integer, exprsInnerLength::Integer,
# Start kernel for each expression to ensure that no warp is working on different expressions
@inbounds Threads.@threads for i in 1:numExprs # multithreaded to speedup dispatching (seems to have improved performance)
numThreads = min(variableColumns, 256)
numThreads = min(variableColumns, 121)
numBlocks = cld(variableColumns, numThreads)
@cuda threads=numThreads blocks=numBlocks fastmath=true interpret_expression(cudaExprs, cudaVars, cudaParams, cudaResults, cudaStepsize, i)