Pages:
Author

Topic: CCminer(SP-MOD) Modded NVIDIA Maxwell / Pascal kernels. - page 32. (Read 2347601 times)

sr. member
Activity: 954
Merit: 250
Bitcore sp-mod #9 has been sendt to the donators.

This time with a +5-10% increase from bitcore sp-mod #8

Bitcore sp-mod #9 is around 50% faster than ccminer 2.2.5

Donate and get access to my private bins. Pm for info.

0: 1080ti: 80% tdp
1: 1070ti: 80% tdp
2: 1070ti: 80% tdp
3: 1070: 80% tdp
4: 1060 3gb: 80% tdp

with more oc and power 40MHASH++ is possible on the 1080ti


I get it, thanks. With this speed BTX hits the top of WTM by far!
sp_
legendary
Activity: 2926
Merit: 1087
Team Black developer
Bitcore sp-mod #9 has been sendt to the donators.

This time with a +5-10% increase from bitcore sp-mod #8

Bitcore sp-mod #9 is around 50% faster than ccminer 2.2.5

Donate and get access to my private bins. Pm for info.

0: 1080ti: 80% tdp
1: 1070ti: 80% tdp
2: 1070ti: 80% tdp
3: 1070: 80% tdp
4: 1060 3gb: 80% tdp

with more oc and power 40MHASH++ is possible on the 1080ti

jr. member
Activity: 94
Merit: 1
I submitted another fix in the makefile, please try again.

now ok, thanks!
sp_
legendary
Activity: 2926
Merit: 1087
Team Black developer
sp_
legendary
Activity: 2926
Merit: 1087
Team Black developer
I submitted another fix in the makefile, please try again.
jr. member
Activity: 94
Merit: 1

I have tried to correct the makefile @ github please try again.


very close already, but not yet...

Code:
ptxas info    : Used 112 registers, 336 bytes cmem[0]
g++  -O3 -march=native -D_REENTRANT -falign-functions=16 -falign-jumps=16 -falign-labels=16 -pthread -L/usr/local/cuda/lib64 -ldl   -o ccminer ccminer-crc32.o ccminer-hefty1.o ccminer-ccminer.o ccminer-pools.o ccminer-util.o ccminer-bench.o ccminer-bignum.o ccminer-api.o ccminer-hashlog.o ccminer-nvml.o ccminer-stats.o ccminer-sysinfos.o ccminer-cuda.o ccminer-nvsettings.o equi/ccminer-equi-stratum.o equi/ccminer-equi.o equi/blake2/ccminer-blake2bx.o equi/ccminer-equihash.o equi/cuda_equi.o heavy/heavy.o heavy/cuda_blake512.o heavy/cuda_combine.o heavy/cuda_groestl512.o heavy/cuda_hefty1.o heavy/cuda_keccak512.o heavy/cuda_sha256.o heavy/bastion.o heavy/cuda_bastion.o ccminer-fuguecoin.o Algo256/cuda_fugue256.o sph/ccminer-fugue.o ccminer-groestlcoin.o cuda_groestlcoin.o ccminer-myriadgroestl.o cuda_myriadgroestl.o lyra2/ccminer-Lyra2.o lyra2/ccminer-Sponge.o lyra2/lyra2RE.o lyra2/cuda_lyra2.o lyra2/lyra2REv2.o lyra2/cuda_lyra2v2.o lyra2/ccminer-Lyra2Z.o lyra2/lyra2Z.o lyra2/cuda_lyra2Z.o Algo256/cuda_bmw256.o Algo256/cuda_cubehash256.o Algo256/cuda_blake256.o Algo256/cuda_groestl256.o Algo256/cuda_keccak256_sm3.o Algo256/cuda_keccak256.o Algo256/cuda_skein256.o Algo256/blake256.o Algo256/decred.o Algo256/vanilla.o Algo256/keccak256.o Algo256/blake2s.o sph/ccminer-blake2s.o Algo256/bmw.o Algo256/cuda_bmw.o crypto/ccminer-xmr-rpc.o crypto/ccminer-wildkeccak-cpu.o crypto/wildkeccak.o crypto/cryptolight.o crypto/cryptolight-core.o crypto/ccminer-cryptolight-cpu.o crypto/cryptonight.o crypto/cryptonight-core.o crypto/cryptonight-extra.o crypto/ccminer-cryptonight-cpu.o crypto/ccminer-oaes_lib.o crypto/ccminer-aesb.o crypto/cpu/ccminer-c_keccak.o JHA/jha.o JHA/jackpotcoin.o JHA/cuda_jha_keccak512.o JHA/cuda_jha_compactionTest.o cuda_checkhash.o quark/cuda_jh512.o quark/cuda_quark_blake512.o quark/cuda_quark_groestl512.o quark/cuda_skein512.o quark/cuda_bmw512.o quark/cuda_quark_keccak512.o quark/nist5.o quark/quarkcoin.o quark/cuda_quark_compactionTest.o neoscrypt/ccminer-neoscrypt.o neoscrypt/ccminer-neoscrypt-cpu.o neoscrypt/cuda_neoscrypt.o pentablake.o skein.o cuda_skeincoin.o ccminer-skein2.o zr5.o skunk/skunk.o skunk/cuda_skunk.o skunk/cuda_skunk_streebog.o sha256/sha256d.o sha256/cuda_sha256d.o sha256/sha256t.o sha256/cuda_sha256t.o sia/sia.o sia/ccminer-sia-rpc.o sph/ccminer-blake2b.o sph/ccminer-bmw.o sph/ccminer-blake.o sph/ccminer-groestl.o sph/ccminer-jh.o sph/ccminer-keccak.o sph/ccminer-skein.o sph/ccminer-cubehash.o sph/ccminer-echo.o sph/ccminer-luffa.o sph/ccminer-sha2.o sph/ccminer-shavite.o sph/ccminer-simd.o sph/ccminer-hamsi.o sph/ccminer-hamsi_helper.o sph/ccminer-streebog.o sph/ccminer-shabal.o sph/ccminer-whirlpool.o sph/ccminer-sha2big.o sph/ccminer-haval.o sph/ccminer-ripemd.o sph/ccminer-sph_sha2.o polytimos.o lbry/lbry.o lbry/cuda_sha256_lbry.o lbry/cuda_sha512_lbry.o lbry/cuda_lbry_merged.o qubit/qubit.o qubit/qubit_luffa512.o qubit/qubit_luffa512_alexis.o qubit/deep.o qubit/luffa.o tribus/tribus.o tribus/cuda_echo512_final.o x11/x11.o x11/fresh.o x11/cuda_x11_luffa512.o x11/cuda_x11_cubehash512.o x11/cuda_x11_cubehash512_80.o x11/cuda_x11_shavite512.o x11/cuda_x11_shavite512_sp.o x11/cuda_x11_simd512.o x11/cuda_x11_echo.o x11/cuda_x11_echo_sp.o x11/cuda_x11_luffa512_Cubehash.o x11/x11evo.o x11/timetravel.o x11/bitcore.o x13/x13.o x13/cuda_x13_hamsi512.o x13/cuda_x13_hamsi512_sp.o x13/cuda_x13_fugue512.o x13/cuda_x13_fugue512_sp.o x13/hsr.o x13/cuda_hsr_sm3.o x13/ccminer-sm3.o x15/x14.o x15/x15.o x15/cuda_x14_shabal512.o x15/cuda_x14_shabal512_sp.o x15/cuda_x15_whirlpool.o x15/whirlpool.o x15/cuda_x15_whirlpool_sm3.o x16r/x16s.o x16r/x16r.o x16r/cuda_x16_echo512.o x16r/cuda_x16_fugue512.o x16r/cuda_x16_shabal512.o x16r/cuda_x16_simd512_80.o x17/x17.o x17/hmq17.o x17/cuda_x17_haval256.o x17/cuda_x17_sha512.o x11/phi.o x11/cuda_streebog_maxwell.o x11/c11.o x11/s3.o x11/sib.o x11/veltor.o x11/cuda_streebog.o ccminer-scrypt.o ccminer-scrypt-jane.o scrypt/blake.o scrypt/keccak.o scrypt/sha256.o scrypt/salsa_kernel.o scrypt/test_kernel.o scrypt/fermi_kernel.o scrypt/kepler_kernel.o scrypt/nv_kernel.o scrypt/nv_kernel2.o scrypt/titan_kernel.o  -lcurl compat/jansson/libjansson.a -lpthread -lcudart -fopenmp -lcrypto -lssl -lz  -ldl  -lcuda -lcrypto -lssl -lz
x16r/x16s.o: In function `scanhash_x16s':
tmpxft_0000648d_00000000-5_x16s.cudafe1.cpp:(.text+0x135f): undefined reference to `x16_shavite512_cpu_hash_80(int, unsigned int, unsigned int, unsigned int*, int)'
tmpxft_0000648d_00000000-5_x16s.cudafe1.cpp:(.text+0x152c): undefined reference to `x16_shavite512_setBlock_80(void*)'
x16r/x16r.o: In function `scanhash_x16r':
tmpxft_000064b7_00000000-5_x16r.cudafe1.cpp:(.text+0x139f): undefined reference to `x16_shavite512_cpu_hash_80(int, unsigned int, unsigned int, unsigned int*, int)'
tmpxft_000064b7_00000000-5_x16r.cudafe1.cpp:(.text+0x15ab): undefined reference to `x16_shavite512_setBlock_80(void*)'
collect2: error: ld returned 1 exit status
Makefile:1241: recipe for target 'ccminer' failed

member
Activity: 312
Merit: 10
is there spmod for lux?
sp_
legendary
Activity: 2926
Merit: 1087
Team Black developer
Code:
mv -f sph/.deps/ccminer-sph_sha2.Tpo sph/.deps/ccminer-sph_sha2.Po
make[2]: *** No rule to make target 'x11/cuda_x11_shavite512_alexis.cu', needed by 'x11/cuda_x11_shavite512_alexis.o'.  Stop.
make[2]: *** Waiting for unfinished jobs....

I have tried to correct the makefile @ github please try again.
sp_
legendary
Activity: 2926
Merit: 1087
Team Black developer
To me it seems like in the latest binary x16s still broken.

fixed now and new exe uploaded.
jr. member
Activity: 94
Merit: 1
X16R/X16S spmod-git #7 has been released

https://github.com/sp-hash/suprminer/releases

-fixed the X16S hash
-faster X16S/X16R (optimizations to hamsi/Echo/echo-simd)



I'm trying to compile that under Ubuntu 18.04 Cuda 9.2 with no success yet

I corrected typo "false" in util.cpp already, but now I have that:

Code:
mv -f sph/.deps/ccminer-sph_sha2.Tpo sph/.deps/ccminer-sph_sha2.Po
make[2]: *** No rule to make target 'x11/cuda_x11_shavite512_alexis.cu', needed by 'x11/cuda_x11_shavite512_alexis.o'.  Stop.
make[2]: *** Waiting for unfinished jobs....

could you correct that? or better check building binaries under linux ?
newbie
Activity: 80
Merit: 0
To me it seems like in the latest binary x16s still broken.
sp_
legendary
Activity: 2926
Merit: 1087
Team Black developer
X16R/X16S spmod-git #7 has been released

https://github.com/sp-hash/suprminer/releases

-fixed the X16S hash
-faster X16S/X16R (optimizations to hamsi/Echo/echo-simd)
sp_
legendary
Activity: 2926
Merit: 1087
Team Black developer
pushed a 17% faster hamsi @ github. 10% faster simd-echo, X16R a few percent faster + fixed X16S
legendary
Activity: 1106
Merit: 1014
any other ideas?
Your bot is stupid, it's copying and pasting other users' messages from/in totally inappropriate topics, so it's obvious that it's a bot posting and not a real user.
legendary
Activity: 2716
Merit: 1094
Black Belt Developer
1x1080ti 2018 = 5x750ti 2014.

Cost to build a "rig", earnings, power consumption, everything. Well except those free pci-e slots you theoretically have today.

That might be true for some algo (some memory intensive ones), but it certainly isn't true for those that make little use of ram. Pascal is 16 nm and Maxwell is 28, that means more transistors and less power.
hero member
Activity: 1151
Merit: 528
1x1080ti 2018 = 5x750ti 2014.

Cost to build a "rig", earnings, power consumption, everything. Well except those free pci-e slots you theoretically have today.

Still too much hope and hype here, is it just me or is BTC about to be programmed to get parabolic again?



I think we sink until 2019 for a variety of TA reasons - we COULD break out before then though.
legendary
Activity: 1176
Merit: 1015
1x1080ti 2018 = 5x750ti 2014.

Cost to build a "rig", earnings, power consumption, everything. Well except those free pci-e slots you theoretically have today.

Still too much hope and hype here, is it just me or is BTC about to be programmed to get parabolic again?


sp_
legendary
Activity: 2926
Merit: 1087
Team Black developer
Still profitable..

in 2014 When bitcoin tanked from $800 to $200 (Us gov. Silkroad dump) the mining profit was destroyed.

"when they can't control the market they attack the POW"
"when they can't control the POW they attack the market"
legendary
Activity: 2716
Merit: 1094
Black Belt Developer
Of course. Making money can't last forever.
legendary
Activity: 1764
Merit: 1024
It modifies you memory timings so it works for everything relevant to that like Monero or other memory intensive stuff.

Interesting, basically all the good algos for Nvidia are GPU based so I never really looked into it, although some of the top earning algos right now are AMD centric (like Dagger), unless you know niches. However, even though I have a dozen or so coins that are off radar, they're all pretty much as fucked. Also majority of my rigs are 1070s.


Going back to a topic that I've brought up a few times over the years, AMD still looks like the right bet, when things get rough it all equalizes with Ethereum. Even though for awhile the profit/$hardware was higher for Nvidia, that gap has basically turned completely negative for like the last six months. So something like a 580 or Vega56 earns quite a bit more per dollar spent then a 1080ti or especially a 1070ti (which has better hash/$ and efficiency).

Although, right now, no one should be buying hardware (or should've been for the last eight months). It's a interesting predicament mining has always circled back around to.
Pages:
Jump to: