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

benchmarking: updated transpiler to drastically reduce the number of transpilations at the expense of memory usage
Some checks are pending
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
CI / Julia ${{ matrix.version }} - ${{ matrix.os }} - ${{ matrix.arch }} - ${{ github.event_name }} (x64, ubuntu-latest, pre) (push) Waiting to run
Details

Browse Source
This commit is contained in:
Daniel
2025-05-19 11:39:49 +02:00
parent 33e7edd4c8
commit f33551e25f
4 changed files with 48 additions and 69 deletions
Download Patch File Download Diff File Expand all files Collapse all files

20
package/src/ExpressionExecutorCuda.jl
View File

@ -49,19 +49,26 @@ end
# Convert Expressions to PTX Code and execute that instead
function evaluate_gpu(expressions::Vector{Expr}, X::Matrix{Float32}, p::Vector{Vector{Float32}}; repetitions=1)::Matrix{Float32}
@assert axes(expressions) == axes(p)
variableCols = size(X, 2)
variableRows = size(X, 1)
numVariableSets = size(X, 2) # nr. of columns of X
variableSetSize = size(X, 1) # nr. of rows of X
variables = CuArray(X)
exprs = Vector{ExpressionProcessing.PostfixType}(undef, length(expressions))
largestParameterSetSize = Utils.get_max_inner_length(p) # parameters get transformed into matrix. Will be nr. of rows in parameter matrix
compiledKernels = Vector{CuFunction}(undef, length(expressions))
kernelName = "evaluate_gpu"
@inbounds Threads.@threads for i in eachindex(expressions)
exprs[i] = ExpressionProcessing.expr_to_postfix(expressions[i])
ex = ExpressionProcessing.expr_to_postfix(expressions[i])
ptxKernel = Transpiler.transpile(ex, variableSetSize, largestParameterSetSize, numVariableSets, i-1, kernelName) # i-1 because julia is 1-based but PTX needs 0-based indexing
compiledKernels[i] = Transpiler.CompileKernel(ptxKernel, kernelName)
end
results = Matrix{Float32}(undef, variableCols, length(exprs))
results = Matrix{Float32}(undef, numVariableSets, length(exprs))
for i in 1:repetitions # Simulate parameter tuning -> local search (X remains the same, p gets changed in small steps and must be performed sequentially, which it is with this impl)
results = Transpiler.evaluate(exprs, variables, variableCols, variableRows, p)
# evaluate
# results = Transpiler.evaluate(exprs, variables, numVariableSets, variableSetSize, p)
results = Transpiler.evaluate(compiledKernels, variables, variableSetSize, p)
end
return results
@ -103,7 +110,6 @@ function interpret_cpu(exprs::Vector{Expr}, X::Matrix{Float32}, p::Vector{Vector
res
end
# Flow
# input: Vector expr == expressions contains eg. 4 expressions
# Matrix X == |expr| columns, n rows. n == number of variabls x1..xn; n is the same for all expressions --- WRONG