floraachy
/
OpenBLAS
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Merge pull request #4354 from imaginationtech/img-rvv-kernel-generator 

[RISC-V] Improve RVV kernel generator LMUL usage
This commit is contained in:
Martin Kroeker 2024-01-17 15:19:37 +01:00 committed by GitHub
parent e3508d3713 4a12cf53ec
commit 88e994116c
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 13 additions and 10 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

21
kernel/riscv64/generate_kernel.py
View File

@ -197,13 +197,13 @@ def generate_gemm_kernel_inner_complex( settings, dest, M, N, vlen, a_regs ):
dest.write("ai += {M}*2;") dest.write("ai += {M}*2;")
dest.write() dest.write()
# for each vector register loaded from matrix A, we require N registers to hold vector-scalar multiply-accumulate results
accumulation_regs = a_regs * N * settings['LMUL_ACC'].value accumulation_regs = a_regs * N
dest.write("// {a_regs} vector regs to hold A array contents, {accumulation_regs} regs to hold values accumulated over k", dest.write("// {a_regs} vector regs to hold A array contents, {accumulation_regs} regs to hold values accumulated over k",
a_regs=a_regs*2, accumulation_regs=accumulation_regs*2 a_regs=a_regs*2, accumulation_regs=accumulation_regs*2
) )
pass_regs = (accumulation_regs + a_regs)*2 pass_regs = (accumulation_regs + a_regs)*2
tmp_regs = 32-pass_regs tmp_regs = (32 // settings['LMUL_ACC'].value) - pass_regs
if tmp_regs < 2: if tmp_regs < 2:
raise RuntimeError("Complex kernel would use too many registers!") raise RuntimeError("Complex kernel would use too many registers!")
@ -337,10 +337,12 @@ def generate_gemm_kernel( settings, OUTPUT ):
M = settings['M'].value M = settings['M'].value
N = settings['N'].value N = settings['N'].value
vlenmax = int( settings['reg_width_bits'].value / settings['ELEN_PARAM'].value ) vlenmax = int(settings['reg_width_bits'].value * settings['LMUL_ACC'].value /
settings['ELEN_PARAM'].value)
a_regs = max(int(M/vlenmax), 1) a_regs = max(int(M/vlenmax), 1)
accumulation_regs = a_regs * N * settings['LMUL_ACC'].value # for each vector register loaded from matrix A, we require N registers to hold vector-scalar multiply-accumulate results
accumulation_regs = a_regs * N
required_regs = accumulation_regs + a_regs required_regs = accumulation_regs + a_regs
if is_complex: if is_complex:
required_regs = required_regs * 2 + 2 required_regs = required_regs * 2 + 2
@ -380,9 +382,9 @@ def generate_gemm_kernel( settings, OUTPUT ):
'''.format(tail_policy=settings['tail_policy'].value)) '''.format(tail_policy=settings['tail_policy'].value))
if required_regs > 32: if required_regs > (32 // settings['LMUL_ACC'].value):
raise Exception("{} vector registers needed during accumulation for unrolling {} x {}{} but only 32 are available".format( raise Exception("{} vector registers needed during accumulation for unrolling {} x {}{} but only {} are available".format(
required_regs, N, M, (" with wide accumulator" if settings['LMUL_ACC'].value > 1 else '') required_regs, N, M, (" with wide accumulator" if settings['LMUL_ACC'].value > 1 else ''), 32 // settings['LMUL_ACC'].value
)) ))
TRMM = (settings['op'].value == 'trmm') TRMM = (settings['op'].value == 'trmm')
@ -448,7 +450,8 @@ def generate_gemm_kernel( settings, OUTPUT ):
def generate_M_tails( dest, settings, M, N ): def generate_M_tails( dest, settings, M, N ):
M_tail = int(M/2) M_tail = int(M/2)
M_tail_min = settings['M_tail_scalar_from'].value M_tail_min = settings['M_tail_scalar_from'].value
vlenmax = int( settings['reg_width_bits'].value / settings['ELEN_PARAM'].value ) vlenmax = int(settings['reg_width_bits'].value * settings['LMUL_ACC'].value
/ settings['ELEN_PARAM'].value )
TRMM = (settings['op'].value == 'trmm') TRMM = (settings['op'].value == 'trmm')
is_complex = settings['complex'].value is_complex = settings['complex'].value
generate_gemm_kernel_inner = generate_gemm_kernel_inner_complex if is_complex else generate_gemm_kernel_inner_real generate_gemm_kernel_inner = generate_gemm_kernel_inner_complex if is_complex else generate_gemm_kernel_inner_real