I am also running 135 puzzles, and using bsgs to run some old lost mining addresses. I have been running for more than half a year in total, but I haven't earned a penny. Haha, I wish you good luck.
Are you using BSGS random to break a 256-bit Bitcoin address where the length of a private key is 64 characters long?

That's wasted energy and resources. Even puzzle 135 is almost impossible to crack by using BSGS random where the prv key is having only 34 chars in length.
Good luck though, I'm currently scanning puzzle 135, bsgs random with 260 exakeys/s and almost giving up.
[+] Random mode
[+] Stats output every 1 seconds
[+] Mode BSGS random
[+] Opening file 135.txt
[+] Added 1 points from file
[+] Range
[+] -- from : 0x4000000000000000000000000000000000
[+] -- to : 0x4fffffffffffffffffffffffffffffffff
[+] N = 0x1000000000000
[+] Bloom filter for 274877906944 elements : 942249.56 MB
[+] Bloom filter for 8589934592 elements : 29445.30 MB
[+] Bloom filter for 268435456 elements : 920.17 MB
[+] Allocating 4096.00 MB for 268435456 bP Points
[+] Reading bloom filter from file keyhunt_bsgs_4_274877906944.blm .... Done!
[+] Reading bloom filter from file keyhunt_bsgs_6_8589934592.blm .... Done!
[+] Reading bP Table from file keyhunt_bsgs_2_268435456.tbl .... Done!
[+] Reading bloom filter from file keyhunt_bsgs_7_268435456.blm .... Done!
[+] Thread 0x4d0c36d8ff9468cc14ad1e038410891c47 conds: ~260 Ekeys/s (260967284607849559562 keys/s)
[/quote]
The 'Stats output every 1 seconds' give you wrong speed, to know the actual speed set the output to at least 10 seconds ( -s 10)
[/quote]
Nope, the speed is correct, changing to -s 10 or even -q doesn't influence the speed too much .
[/quote]
Totally agree, #135 using BSGS will take million years to crack. I went further and changes JLP vanitysearch, that it will calculate 2^30 precomputed table and run that in each thread against public key using bloomfilter, using symmetry and 8x 4090 can do 48 billion of 2^31 ranges per seacond, and that will take millions of years to go through 135 bits. I think BSGS at this stage is useless.
[/quote]
So if your 8x 4090 can do 48 billion of 2^31 ranges per second (equal with 103,079,215,104,000,000,000 keys per second) it means your hardware is X2 slower than an AMD EPYC 64 cores + 1T RAM , running on BSGS random, doing 260,949,667,359,921,111,308 keys per second. Indeed, random is random, can never hit the private key especially on bigger ranges, while sequential search is a guarantee.