Tanguy Pruvot
56ed0fed05
blake: remove hashharder workaround
2014-12-01 05:58:01 +01:00
Tanguy Pruvot
8ad180cc70
various small changes
...
heavy: reduce by 256 threads default intensity to all -i 20
cuda: put static thread init bools outside the code (made once)
api: fix nvml header to build without
2014-11-28 20:57:35 +01:00
Tanguy Pruvot
1b65cd05cc
heavy: add error checks, fix strict aliasing and linux
...
The core problem was the cuda hefty Thread per block set to high
but took me several hours to find that...
btw... +25% in heavy 12500 with 256 threads per block... vs 128 & 512
if max reg count is set to 80...
2014-11-27 09:14:59 +01:00
Tanguy Pruvot
6ae28162db
various extern cleanup + api history uids and gpu SM
...
uids could be useful to create graphes from history data
Note: please do a clean build after this commit (changes in miner.h)
2014-11-26 11:55:42 +01:00
Tanguy Pruvot
73f22b237a
Prepare trap of hardware/mem failures
2014-11-20 18:44:25 +01:00
Tanguy Pruvot
b4ef7b981f
scan range: add boundary check, cant be > UINT32_MAX
2014-11-17 16:51:36 +01:00
Tanguy Pruvot
438308b3a2
Rework benchmark mode and min/max range
...
Was maybe my fault, but the benchmark mode was
always recomputing from nonce 0.
Also fix blake if -d 1 is used (one thread but second gpu)
stats: do not use thread id as key, prefer gpu id...
2014-11-16 23:28:18 +01:00
Tanguy Pruvot
b128312efb
cuda: store device SM in a global var
...
sample usage made for blake and fugue (higher intensity for SM5.2)
add these to cuda_helper and clean unused code
2014-11-11 19:11:16 +01:00
Tanguy Pruvot
11c5ec810d
Handle intensity param in all algos
...
and add a check related to start/max nounce params
2014-11-09 22:27:32 +01:00
Tanguy Pruvot
4c3964539f
Fix vc debug builds, missing symbols
2014-11-06 17:42:01 +01:00
Tanguy Pruvot
12fafd5687
Try to reconnect on pool duplicates
...
reduce log announces and define uchar in miner.h
2014-11-04 15:14:24 +01:00
Tanguy Pruvot
187e293f71
blake: some fine tuning + cleanup
2014-11-03 20:55:03 +01:00
Tanguy Pruvot
5bc969fa57
Some work on data alignment
...
linux: add -march=native (we build it ourself) and some other flags
+ remove unused vars (seen with -Wall)
2014-11-03 16:40:13 +01:00
Tanguy Pruvot
93bb428bdf
blake: rewrite the cache system
...
Unlike other hash algos, blake256 compute the hash
with blocks of 64 bytes.
We can do the first part on the cpu, only the 4 last int32
are computed on gpu (including the tested nonce)
Previous method was also using this kind of cache with a crc.
Blake Hash Speed: +5%
2014-11-03 16:33:59 +01:00
Tanguy Pruvot
ba33492592
blake: return to ptarget 6:7 compare
...
clz can be erroneous, ex 0xE0 vs 0xF0
2014-09-19 05:01:16 +02:00
Tanguy Pruvot
91eea0d76b
blake: remove int cudaMemcpyToSymbol for MSVC
...
use clz (leading zeros) asm func for a fast gpu compare of ptarget[6]:[7]
add also missing windows ctz/clz host functions
New NEOS speed: 227MH to 270MH (Gigabyte 750Ti Black Edition)
2014-09-13 17:31:01 +02:00
Tanguy Pruvot
9efe0b965d
blake: only use high part of target on gpu
...
Add another few MH/s boost :)
2014-09-13 00:15:34 +02:00
Tanguy Pruvot
8925a7551f
blake: final cleanup (225MH/s)
2014-09-11 20:16:16 +02:00
Tanguy Pruvot
347d4e4928
blake: +8MH/s on linux, weird optimisation
...
Like doom/luffa, using a int pos make the proc faster
2014-09-11 02:33:34 +02:00
Tanguy Pruvot
cec5baea95
enable colors by default, except for syslog
...
debug: show compared hash diffs in color
2014-09-09 21:59:03 +02:00
Tanguy Pruvot
402e416853
Add pentablake algo (-a penta)
...
Signed-off-by: Tanguy Pruvot <tanguy.pruvot@gmail.com>
2014-09-09 21:58:58 +02:00
Tanguy Pruvot
42eafcbe85
Put CRC-32 function in a new unit
...
Signed-off-by: Tanguy Pruvot <tanguy.pruvot@gmail.com>
2014-09-06 19:41:48 +02:00
Tanguy Pruvot
5ccd166916
blake: introduce pdata head cache (speed x2)
2014-09-06 16:26:53 +02:00
Tanguy Pruvot
65909ec3b7
blake: handle case when 2 hashes are found in a call
2014-09-06 11:43:32 +02:00
Tanguy Pruvot
52ec8830b1
blake: blakecoin variant now works
2014-09-06 01:21:30 +02:00
Tanguy Pruvot
ecc86af102
blake: sometimes faster, or not
2014-09-05 21:12:38 +02:00
Tanguy Pruvot
3356e6f8bf
blake: some more KH/s on linux
2014-09-05 20:00:07 +02:00
Tanguy Pruvot
12fefe5de0
blake: add a few more MH/s, prepare blakecoin
2014-09-05 19:12:29 +02:00
Tanguy Pruvot
5682b7d241
blake: add also blakecoin (8-rounds) variant
2014-09-05 19:12:29 +02:00
Tanguy Pruvot
e1159629b4
blake: typo for windows on last commit
2014-09-05 09:47:24 +02:00
Tanguy Pruvot
746398f435
blake: fix reduced speed on windows, wtf
2014-09-04 17:34:30 +02:00
Tanguy Pruvot
3341e0324f
blake: speed +10%, no more size conversions
2014-09-04 16:17:11 +02:00
Tanguy Pruvot
806c3e8691
enhance double scan checks
2014-09-04 11:11:39 +02:00
Tanguy Pruvot
69616b37ac
hashlog: prepare store of scanned range
2014-09-03 14:49:13 +02:00
Tanguy Pruvot
1a4391d7ff
hashlog: prevent double computing on jobs already done
2014-09-03 12:20:24 +02:00
Tanguy Pruvot
049e577301
tmp blake log
2014-09-03 09:49:14 +02:00
Tanguy Pruvot
43d3e93e1a
blake: set a max throughput
2014-09-03 09:29:51 +02:00
Tanguy Pruvot
7e595a36ea
blake: cleanup, remove d_hash buf, not in a chain
...
host: only bencode if gpu hash was found
2014-09-03 08:33:55 +02:00
Tanguy Pruvot
de80c7e9d1
blake: remove unused parameter and fix index in d_hash
...
that reduce the speed to 92MH/s but the next commit
give us 30 more
so, todo: merge the whole checkhash proc in gpu_hash
and remove this d_hash buffer...
2014-09-02 20:20:53 +02:00
Tanguy Pruvot
530732458a
blake: use a constant for threads, reduce mallocated d_hash size
...
and clean a bit more...
2014-09-01 20:00:25 +02:00
Tanguy Pruvot
0aeac878ef
blake: tune up and cleanup, ~100 MH/s on a normal 750Ti
...
tested on linux and windows (x86 binary)...
but there is a high number of duplicated shares... weird
2014-09-01 12:24:26 +02:00
Tanguy Pruvot
c17d11e377
add "blake" 256, 14 rounds (for NEOS blake, not BlakeCoin)
...
also remove "missing" file, its old and not compatible with ubuntu 14.04
2014-09-01 05:53:36 +02:00
