Issue with v1.3 #3
Description
This happens on non-SSD machine (Ubuntu 24.04, ext4 encrypted via cryptsetup):
$ time ./Mark1 --range 18889465931478580854783:37778931862957161709567 --pubkey 03726b574f193e374686d8e12bc6e4142adeb06770e0a2856f5e4ad89f66044755 --dp_point 200000000 --dp_bits 8 --ram 32
=========== Phase-0: Data summary ==========
DP table (SSD): 20.0Gb ( 200000000 / 536870912 slots, load 37.25% )
Bloom (RAM): 1.86Gb
========== Phase-1: Building traps =========
Unique traps: 200000000/200000000 (done)
This is where it freezes. Strace reporting no activity within the process:
$ sudo strace -p 35243
strace: Process 35243 attached
The process is still visible in btop with 21G memory usage, 49 threads and 0.0% cpu usage. It is impossible to kill the process with sudo kill -9 35243, the process is still sitting in the queue.
Would it be possible to have an option for disabling SSD features and keep all the data needed in the memory?
update:
I created the RAM disk of 25G and it is now running:
$ time ./Mark1 --range 18889465931478580854783:37778931862957161709567 --pubkey 03726b574f193e374686d8e12bc6e4142adeb06770e0a2856f5e4ad89f66044755 --dp_point 200000000 --dp_bits 8 --ram 32
=========== Phase-0: Data summary ==========
DP table (SSD): 20.0Gb ( 200000000 / 536870912 slots, load 37.25% )
Bloom (RAM): 1.86Gb
========== Phase-1: Building traps =========
Unique traps: 200000000/200000000 (done)
=========== Phase-2: Kangaroos =============
Speed: 104.35 MH/s | Hops: 89462669312 | Restart wild: 0 | Time: 0:14:15
============= Phase-3: Result ==============
Private key : 0x0000000000000000000000000000000000000000000004C5CE114686A1336E07
Found by thr: 28
Wild wraps : 0 [no wrap]
Wild restart: 0
Total time : 00:14:15.893
Private key : saved to FOUND.txt
real 22m25.954s
user 1066m26.086s
sys 0m25.369s
It is slower than v1.0: 21:52.683 vs 22:25:954 ~4.6%
Running from the ramdisk: (process has 96 threads and takes 21G of RAM)
$ time ./Mark1 --range 18889465931478580854783:37778931862957161709567 --pubkey 03726b574f193e374686d8e12bc6e4142adeb06770e0a2856f5e4ad89f66044755 --dp_point 200000000 --dp_bits 8 --ram 32 --load-dp dp_table.bin
=========== Phase-0: Data summary ==========
DP table (SSD): 20.0Gb ( 200000000 / 536870912 slots, load 37.25% )
Bloom (RAM): 1.86Gb
========== Phase-1: Building traps =========
Loaded 536870912 traps (done)
=========== Phase-2: Kangaroos =============
Speed: 104.67 MH/s | Hops: 109282066432 | Restart wild: 0 | Time: 0:17:25^C
real 18m36.350s
user 839m27.328s
sys 0m11.352s
I killed it, the number of hops went above what I observed in the succesful run: 109282066432 > 89462669312; time is also beyond the baseline.
Running from the non-SSD disk:
It was slow to read the data from the encrypted ext4 volume (samples every second):
Device tps kB_read/s kB_wrtn/s kB_dscd/s kB_read kB_wrtn kB_dscd
...
dm-0 626.00 2492.00 12.00 0.00 2492 12 0
dm-0 771.00 7176.00 0.00 0.00 7176 0 0
dm-0 699.00 2796.00 0.00 0.00 2796 0 0
dm-0 1034.00 8228.00 0.00 0.00 8228 0 0
dm-0 651.00 2604.00 0.00 0.00 2604 0 0
dm-0 707.00 2816.00 12.00 0.00 2816 12 0
dm-0 1091.00 8452.00 0.00 0.00 8452 0 0
dm-0 757.00 3028.00 0.00 0.00 3028 0 0
dm-0 1116.00 8516.00 0.00 0.00 8516 0 0
dm-0 1350.00 9428.00 0.00 0.00 9428 0 0
dm-0 1585.00 10340.00 12.00 0.00 10340 12 0
dm-0 1642.00 10632.00 0.00 0.00 10632 0 0
...
$ time ./Mark1 --range 18889465931478580854783:37778931862957161709567 --pubkey 03726b574f193e374686d8e12bc6e4142adeb06770e0a2856f5e4ad89f66044755 --dp_point 200000000 --dp_bits 8 --ram 32 --load-dp dp_table.bin
=========== Phase-0: Data summary ==========
DP table (SSD): 20.0Gb ( 200000000 / 536870912 slots, load 37.25% )
Bloom (RAM): 1.86Gb
========== Phase-1: Building traps =========
Loaded 198180864^C
real 42m14.053s
user 0m29.460s
sys 0m46.856s