Lerking/OSACA
1
0
Fork 0
mirror of https://github.com/RRZE-HPC/OSACA.git synced 2026-01-05 02:30:08 +01:00
Code Issues Packages Projects Releases Wiki Activity

initial upload

Browse Source
This commit is contained in:
Jan Laukemann
2017-07-17 15:29:56 +02:00
parent 183802b110
commit a1dc3b639b
67 changed files with 6265 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

65
testcases/vunpckhpd-xmmxmmxmm.S Normal file
View File

@@ -0,0 +1,65 @@
#define INSTR vunpckhpd
#define NINST 24
#define N edi
#define i r8d
.intel_syntax noprefix
.globl ninst
.data
ninst:
.long NINST
.text
.globl latency
.type latency, @function
.align 32
latency:
push rbp
mov rbp, rsp
xor i, i
test N, N
jle done
# create DP 1.0
vpcmpeqw xmm0, xmm0, xmm0 # all ones
vpsllq xmm0, xmm0, 54 # logical left shift: 11111110..0 (54=64-(10-1))
vpsrlq xmm0, xmm0, 2 # logical right shift: 1 bit for sign; leading mantissa bit is zero
# copy DP 1.0
vmovaps xmm0, xmm0
vmovaps xmm1, xmm0
# Create DP 2.0
vaddpd xmm1, xmm1, xmm1
# Create DP 0.5
vdivpd xmm2, xmm0, xmm1
loop:
inc i
INSTR xmm0, xmm1, xmm0
INSTR xmm1, xmm0, xmm0
INSTR xmm0, xmm1, xmm0
INSTR xmm1, xmm0, xmm0
INSTR xmm0, xmm1, xmm0
INSTR xmm1, xmm0, xmm0
INSTR xmm0, xmm1, xmm0
INSTR xmm1, xmm0, xmm0
INSTR xmm0, xmm1, xmm0
INSTR xmm1, xmm0, xmm0
INSTR xmm0, xmm1, xmm0
INSTR xmm1, xmm0, xmm0
INSTR xmm0, xmm1, xmm0
INSTR xmm1, xmm0, xmm0
INSTR xmm0, xmm1, xmm0
INSTR xmm1, xmm0, xmm0
INSTR xmm0, xmm1, xmm0
INSTR xmm1, xmm0, xmm0
INSTR xmm0, xmm1, xmm0
INSTR xmm1, xmm0, xmm0
INSTR xmm0, xmm1, xmm0
INSTR xmm1, xmm0, xmm0
INSTR xmm0, xmm1, xmm0
INSTR xmm1, xmm0, xmm0
cmp i, N
jl loop
done:
mov rsp, rbp
pop rbp
ret
.size latency, .-latency