笔记1月5 - 技术栈

bash 复制代码

./build/X86_MOESI_hammeryz1wVCBuffer/gem5.opt -d m5out/20250105NoSleepX86_MOESI_hammeryz1wVCBuffer_X86KvmCPUCHKPT/    configs/deprecated/example/fs.py --kernel=/home/yz/.cache/gem5/x86-linux-kernel-4.19.83 --disk=/home/yz/.cache/gem5/x86-parsec      --cpu-type=X86KvmCPU      --num-cpus=64   --ruby   --network=garnet   --topology=Mesh_XY   --mesh-rows=8 --num-dirs=64  --num-l2caches=64  --script=configs/boot/yzhack_back_ckpt.rcS

可能几分钟跑完

bash 复制代码

 ./build/X86_MOESI_hammeryz1wVCBuffer/gem5.opt  -d m5out/used_20250105NoSleepX86_MOESI_hammeryz1wVCBuffer_X86KvmCPUCHKPT/ configs/deprecated/example/fs.py     --checkpoint-restore=1  --checkpoint-dir=/home/yz/myprojects/2024GEM5/parsec-tests/yzmodifiedgem5/m5out/20250105NoSleepX86_MOESI_hammeryz1wVCBuffer_X86KvmCPUCHKPT/ --restore-with-cpu=X86AtomicSimpleCPU   --cpu-type=X86AtomicSimpleCPU     --num-cpus=64   --ruby   --network=garnet   --topology=Mesh_XY   --mesh-rows=8 --num-dirs=64  --num-l2caches=64  --script=configs/yz2023Nov/large/yzjustexit.script

这个调用checkpoint只要几十秒钟，确切一点，在我的10700k上是19秒。

下一步是用kvm的checkpoint 运行cpu-type=X86AtomicSimpleCPU .

bash 复制代码

 ./build/X86_MOESI_hammeryz1wVCBuffer/gem5.opt  -d m5out/used_20250105NoSleepX86_MOESI_hammeryz1wVCBuffer_X86KvmCPUCHKPT/blacksholes configs/deprecated/example/fs.py     --checkpoint-restore=1  --checkpoint-dir=/home/yz/myprojects/2024GEM5/parsec-tests/yzmodifiedgem5/m5out/20250105NoSleepX86_MOESI_hammeryz1wVCBuffer_X86KvmCPUCHKPT/ --restore-with-cpu=X86AtomicSimpleCPU   --cpu-type=X86AtomicSimpleCPU     --num-cpus=64   --ruby   --network=garnet   --topology=Mesh_XY   --mesh-rows=8 --num-dirs=64  --num-l2caches=64  --script=configs/yz2023Nov/large/yzfs_largeparsecblacksholes.script

m5term 那边报错：

"hda: status error: status=0x00 { }

hda: possibly failed opcode: 0xc8

hda: drive not ready for command

hda: status error: status=0x00 { }

hda: possibly failed opcode: 0xc8

hda: drive not ready for command

hda: status error: status=0x00 { }

hda: possibly failed opcode: 0xc8

hda: drive not ready for command

hda: status error: status=0x00 { }

hda: possibly failed opcode: 0xc8

hda: drive not ready for command

ide0: reset: master: error (0x00?)

hda: status error: status=0x00 { }

hda: possibly failed opcode: 0xc8

hda: drive not ready for command

hda: status error: status=0x00 { }

hda: possibly failed opcode: 0xc8

hda: drive not ready for command

hda: status error: status=0x00 { }

hda: possibly failed opcode: 0xc8

hda: drive not ready for command

hda: status error: status=0x00 { }

hda: possibly failed opcode: 0xc8

hda: drive not ready for command

ide0: reset: master: error (0x00?)

print_req_error: I/O error, dev hda, sector 12650752

EXT4-fs error (device hda1): ext4_find_entry:1455: inode #393217: comm bash: reading directory lblock 0

print_req_error: I/O error, dev hda, sector 2048

Buffer I/O error on dev hda1, logical block 0, lost sync page write

EXT4-fs (hda1): I/O error while writing superblock

Aborting journal on device hda1-8.

print_req_error: I/O error, dev hda, sector 21243936

Buffer I/O error on dev hda1, logical block 2655236, lost sync page write

JBD2: Error -5 detected when updating journal superblock for hda1-8.

EXT4-fs (hda1): Remounting filesystem read-only

print_req_error: I/O error, dev hda, sector 12650752

EXT4-fs error (device hda1): ext4_find_entry:1455: inode #393217: comm bash: reading directory lblock 0

print_req_error: I/O error, dev hda, sector 2048

Buffer I/O error on dev hda1, logical block 0, lost sync page write

EXT4-fs (hda1): I/O error while writing superblock

print_req_error: I/O error, dev hda, sector 12650752

EXT4-fs error (device hda1): ext4_find_entry:1455: inode #393217: comm bash: reading directory lblock 0

print_req_error: I/O error, dev hda, sector 2048

Buffer I/O error on dev hda1, logical block 0, lost sync page write

EXT4-fs (hda1): I/O error while writing superblock

/tmp/runscript: line 1: cd: /home/gem5/parsec-benchmark: Input/output error

/tmp/runscript: line 1: env.sh: No such file or directory

/tmp/runscript: line 1: parsecmgmt: command not found

print_req_error: I/O error, dev hda, sector 42257752

/tmp/runscript: line 1: /bin/sleep: Input/output error

再次验证以后，发现just exit能运行，blacksholes不行。所以删除kvm的checkpoint。等6号再来。

7好发现，atomic的checkpoint，跑timingsimple依旧类似错误。

重新用timingsimple存checkpoint。