Takashi Nakamura tanakamura
tanakamura
/ matmulgccfma
Created
October 23, 2014 17:23
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| matmul_x86_fma_: | |
| .LFB1165: | |
| .cfi_startproc | |
| pushq %rbp | |
| .cfi_def_cfa_offset 16 | |
| .cfi_offset 6, -16 | |
| addq %rdi, %rcx | |
| vxorps %xmm0, %xmm0, %xmm0 | |
| movq %rsp, %rbp | |
| .cfi_def_cfa_register 6 |
tanakamura
/ lookup8.c
Last active
August 29, 2015 14:27
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| #include <immintrin.h> | |
| #include <stdio.h> | |
| #include <intrin.h> | |
| /* | |
| lookup8: | |
| subq $12, %rsp | |
| .seh_stackalloc 12 | |
| .seh_endprologue | |
| movl %edx, (%rsp) |
tanakamura
/ median.cpp
Created
December 14, 2015 05:31
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| #include <stdio.h> | |
| #include <string.h> | |
| #include <stdlib.h> | |
| #include <time.h> | |
| double | |
| sec(void) | |
| { | |
| struct timespec ts; |
tanakamura
/ mem.c
Last active
May 6, 2016 00:55
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| /* | |
| ==4KB== | |
| memset L1/L2 1.8[usec], 2230.9[MB/s] | |
| memset L2 1.9[usec], 2198.6[MB/s] | |
| memset L1 5.9[usec], 696.6[MB/s] | |
| memset wo cache 19.9[usec], 206.2[MB/s] | |
| flush wb 10.1[usec], 406.2[MB/s] | |
| flush4k wb 8.9[usec], 459.9[MB/s] | |
| flush empty 4.1[usec], 990.8[MB/s] | |
| flush4k empty 3.2[usec], 1270.1[MB/s] |
tanakamura
/ top.v
Created
May 13, 2016 13:02
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| module top(led_0, | |
| sw_0); | |
| (* IOSTANDARD="LVCMOS33"*) (* PACKAGE_PIN = "T22"*) output led_0; | |
| (* IOSTANDARD="LVCMOS25"*) (* PACKAGE_PIN = "F22"*) input sw_0; | |
| assign led_0 = sw_0; | |
| endmodule |
tanakamura
/ build.tcl
Created
May 13, 2016 13:03
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| set output_dir ./out | |
| file mkdir $output_dir | |
| read_verilog top.v | |
| synth_design -top top -part xc7z020clg484-1 -flatten rebuilt | |
| opt_design | |
| place_design | |
| phys_opt_design |
tanakamura
/ parallellagpio.v
Created
May 18, 2016 09:44
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| module design_1_wrapper | |
| (); | |
| wire [63:0] gpio_out0; | |
| wire [63:0] gpio_in0; | |
| assign gpio_in0[0] = gpio_out0[1]; | |
| PS7 PS7_i ( | |
| .EMIOGPIOO (gpio_out0), |
tanakamura
/ clflushopt.c
Last active
May 20, 2016 06:54
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| /* | |
| clflush 4278.300100[MiB/s] | |
| clflushopt 44717.389375[MiB/s] | |
| memset 30407.684046[MiB/s] | |
| */ | |
| #include <stdio.h> | |
| #include <x86intrin.h> | |
| #include <sys/time.h> | |
| #include <string.h> |
tanakamura
/ knl.txt
Last active
August 3, 2016 05:09
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| (https://github.com/tanakamura/instruction-bench) | |
| Intel(R) Xeon Phi(TM) CPU 7210 @ 1.30GHz | |
| == latency/throughput == | |
| reg64: add: latency: CPI= 1.11, IPC= 0.90 | |
| reg64: add:throughput: CPI= 0.64, IPC= 1.57 | |
| reg64: lea: latency: CPI= 1.03, IPC= 0.97 | |
| reg64: lea:throughput: CPI= 0.65, IPC= 1.55 | |
| reg64: load: latency: CPI= 4.04, IPC= 0.25 | |
| reg64: load:throughput: CPI= 1.02, IPC= 0.98 |
tanakamura
/ knl-binutils.txt
Created
August 3, 2016 07:34
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Performance counter stats for 'sh -c ../configure; make -j256': | |
| 631718.546417 task-clock (msec) # 4.435 CPUs utilized | |
| 79,129 context-switches # 0.125 K/sec | |
| 46,742 cpu-migrations # 0.074 K/sec | |
| 14,459,831 page-faults # 0.023 M/sec | |
| 794,193,966,671 cycles # 1.257 GHz (50.80%) | |
| <not supported> stalled-cycles-frontend | |
| <not supported> stalled-cycles-backend | |
| 396,607,076,972 instructions # 0.50 insns per cycle (75.66%) |