[ANN][XCN] Cryptonite | 1st mini-blockchain coin | M7 PoW | No Premine

[bitfreak!]

My speculations are not "baseless." As I said, I have direct evidence pointing to a day one GPU miner. I do not wish to release this evidence because I think more may come out about this situation over time and I'm happy to let the secret miner make statements contradicted by the evidence, further pointing to deception. And further because doing so may give future clandestine GPU-on-launch miners some ideas on how to avoid detection. However, I might be convinced to share the evidence with a trusted third party, in confidence.

Well my comment wasn't directed at you for a start, but you just made a baseless accusation because you failed to provide any evidence. If you do have some information worth sharing then just release it, don't beat around the bush.

Even without disassembly or decompilation, is it possible the PoW could have leaked? Did you treat the PoW as a tightly guarded secret and not discuss it with anyone nor collaborate with anyone during development?

Only 4 people had access to the source code during development, and I'm very sure none of them leaked it.

[1713547047]

1713547047

[1713547047]

1713547047

[1713547047]

1713547047

[deathmul]

Any GPU miner for AMD?

[jmumich]

I guess you meant it wasn't fair. There was a public beta you could make GPU miner too so don't complain. And I think there is no such thing as fair distribution. People on here are always complaining about unfair distribution/no pools/no windows wallet/premine/instamine/low trading volume/scrypt algo/missing train/etc. There is no perfect coin, man.

Perfection? No, we've reached a point that all of the above are just basic requirements.

If a GPU miner was possible to be created, XCN should have never been launched until the GPU miner was ready. This was a major mistake and a very valid reason to call this an unfair launch IMO.

To give you an example on AWS:

A 32 vCPU core EC2 instance costs ~0.27 $/hr and produces around 1700+ KH/s
A 1 GPU + 8 vCPU core EC2 instance costs ~0.065 $/hr and produces 1800+ KH/s (CPU + GPU mining -- using the non optimized GPU miner)

If you do the math, you realize that it was at least 4-5 times more possible for those who had the private GPU miner on AWS to find a block than the rest, provided that both had spent the same amount of $ to rent instances. Now if we think outside AWS and take into consideration that GPU farms could have been used, the advantage becomes much bigger.

So, please, this is not about a small detail regarding perfection.

PS: Also, since I deleted some messages of mine earlier and since someone had mentioned it, I'd like to say again that I never said anywhere that wolf's optimizations are a problem. 10-15% advantage is not a problem at all. But 400-500% advantage is was a clear possible problem (EDIT: at launch).

The problem with your assumption is that there are far fewer available G2 instances than there are available 32 core CPU instances.  And it ignores 16 core, 8 core, etc, all of which were available at a similar price per core.   Your hypothetical miner could never have run 4-5x as many GPU instances and maintained the same price of the G2 instances. 

The early mining from large miners is consistent with a miner using a large number of Amazon EC2 instances for CPU mining.   

[Youghoor]

Any GPU miner for AMD?

not yet have to wait for it

[Amph]

at least this thing is still profitable, better to mine some before amd guys arrive

[smooth]

Even without disassembly or decompilation, is it possible the PoW could have leaked? Did you treat the PoW as a tightly guarded secret and not discuss it with anyone nor collaborate with anyone during development?

Only 4 people had access to the source code during development, and I'm very sure none of them leaked it.

Was the testnet version stripped of symbols and debug tags (on all OSs)?

Was the PoW design discussed with other people? That could have been enough to put together a rough miner to be tweaked with details upon release. The miner in question did not start mining until approximately block 150. That would have been enough time to finish a miner that was already mostly implemented.

I guess it is also possible that someone with experience and a good library of GPU hash functions and bignum multiplication could have thrown a miner together in 2-3 hours without a head start. That's another possibility I suppose.

[Amph]

Even without disassembly or decompilation, is it possible the PoW could have leaked? Did you treat the PoW as a tightly guarded secret and not discuss it with anyone nor collaborate with anyone during development?

Only 4 people had access to the source code during development, and I'm very sure none of them leaked it.

Was the testnet version stripped of symbols and debug tags (on all OSs)?

Was the PoW design discussed with other people? That could have been enough to put together a rough miner to be tweaked with details upon release. The miner in question did not start mining until approximately block 150. That would have been enough time to finish a miner that was already mostly implemented.

I guess it is also possible that someone with experience and a good library of GPU hash functions and bignum multiplication could have thrown a miner together in 2-3 hours without a head start. That's another possibility I suppose.

was the nethash that big at block 150?

[bitfreak!]

The miner in question did not start mining until approximately block 150. That would have been enough time to finish a miner that was already mostly implemented.

You're obviously referring to the large miner who started mining at block 146. Just because they had such a large amount of hashing power doesn't prove they had a GPU miner. But I agree there is a chance someone got a GPU miner whipped up within 150 blocks, because they had a lot of time to prepare for our release. Either way they've stopped mining quite some time ago now and we have more important issues to focus on.

[smooth]

was the nethash that big at block 150?

Probably not, I don't know, I didn't even start mining at all until later.

The miner in question was still getting a very high percentage of blocks (in some cases several blocks in a row) well into the thousands (i.e. days later) when the hash rate had gone up a lot.

[cubydu]

The miner in question did not start mining until approximately block 150. That would have been enough time to finish a miner that was already mostly implemented.

You're obviously referring to the large miner who started mining at block 146. Just because they had such a large amount of hashing power doesn't prove they had a GPU miner. But I agree there is a chance someone got a GPU miner whipped up within 150 blocks, because they had a lot of time to prepare for our release. Either way they've stopped mining quite some time ago now and we have more important issues to focus on.

Fu...k!  I'm tired of all this. I started mine with 100 c3.8xlarge after 2 hours after launch. After 2 hours because I couldn't make daemon  from src

[discodancer]

The miner in question did not start mining until approximately block 150. That would have been enough time to finish a miner that was already mostly implemented.

You're obviously referring to the large miner who started mining at block 146. Just because they had such a large amount of hashing power doesn't prove they had a GPU miner. But I agree there is a chance someone got a GPU miner whipped up within 150 blocks, because they had a lot of time to prepare for our release. Either way they've stopped mining quite some time ago now and we have more important issues to focus on.

Fu...k!  I'm tired of all this. I started mine with 100 c3.8xlarge after 2 hours after launch. After 2 hours because I couldn't make daemon  from src

c3.8xlarge still profitable?

[cubydu]

The miner in question did not start mining until approximately block 150. That would have been enough time to finish a miner that was already mostly implemented.

You're obviously referring to the large miner who started mining at block 146. Just because they had such a large amount of hashing power doesn't prove they had a GPU miner. But I agree there is a chance someone got a GPU miner whipped up within 150 blocks, because they had a lot of time to prepare for our release. Either way they've stopped mining quite some time ago now and we have more important issues to focus on.

Fu...k!  I'm tired of all this. I started mine with 100 c3.8xlarge after 2 hours after launch. After 2 hours because I couldn't make daemon  from src

c3.8xlarge still profitable?

no, only first 2-3 days

[go6ooo1212]

Im getting about 3050-3080 Kh/s per card with Palit 750TI stormX OC with the new miner. Im with 340.52 Nvidia driver, and this is my command line:
Cryptonite-CudaMiner-compute_35.exe -o stratum+tcp://xcnpool.1gh.com:7333 -u Y -p x -t 6
I've tried with compute_30 but the speed is about 2700kh.
How could someone reach 4MH ?

[myagui]

Im getting about 3050-3080 Kh/s per card with Palit 750TI stormX OC with the new miner. Im with 340.52 Nvidia driver, and this is my command line:
Cryptonite-CudaMiner-compute_35.exe -o stratum+tcp://xcnpool.1gh.com:7333 -u Y -p x -t 6
I've tried with compute_30 but the speed is about 2700kh.
How could someone reach 4MH ?

Check your CPU usage. It's quite likely that your CPU is the bottleneck. If that's the case, you might find that you are better off just mining XCN with 4 cards, and mine something else with the other two (some other algo that doesn't stress the CPU).

EDIT: Then again, the windows build is largely "beta", so I could be completely off... I suppose that the best performance right now is with Linux.
Good Luck
~ Myagui

[go6ooo1212]

Im getting about 3050-3080 Kh/s per card with Palit 750TI stormX OC with the new miner. Im with 340.52 Nvidia driver, and this is my command line:
Cryptonite-CudaMiner-compute_35.exe -o stratum+tcp://xcnpool.1gh.com:7333 -u Y -p x -t 6
I've tried with compute_30 but the speed is about 2700kh.
How could someone reach 4MH ?

Check your CPU usage. It's quite likely that your CPU is the bottleneck. If that's the case, you might find that you are better off just mining XCN with 4 cards, and mine something else with the other two (some other algo that doesn't stress the CPU).

EDIT: Then again, the windows build is largely "beta", so I could be completely off... I suppose that the best performance right now is with Linux.
Good Luck
~ Myagui

Thank you! I've checked the cpu load , and it get between 34-50%. I think its not stressed ,but will try turning off 2 of the cards

[ivcelmik]

Hy guys need help i have 2 Rigs with 6x750ti each ,one with 8Gb Ram and Win7x64 and the second with 6GB Ram an Win 8x64,the first one works perfect and the second stops after 5 ,10 or 30 minutes if u have any ideas or ..

Thx in advance

[Jungian]

Hy guys need help i have 2 Rigs with 6x750ti each ,one with 8Gb Ram and Win7x64 and the second with 6GB Ram an Win 8x64,the first one works perfect and the second stops after 5 ,10 or 30 minutes if u have any ideas or ..

Thx in advance

I've heard that RAM may be a problem. 6gb ram for win8 sounds a bit thin in my ears, but someone more qualified can hopefully answer. I suggest you try in the cudaminer thread https://bitcointalk.org/index.php?topic=167229.0

They may know a little more there.

[djm34]

Im getting about 3050-3080 Kh/s per card with Palit 750TI stormX OC with the new miner. Im with 340.52 Nvidia driver, and this is my command line:
Cryptonite-CudaMiner-compute_35.exe -o stratum+tcp://xcnpool.1gh.com:7333 -u Y -p x -t 6
I've tried with compute_30 but the speed is about 2700kh.
How could someone reach 4MH ?

Actually I can, (my card is OCed), however the 750ti being a second card, it kills at the same time the performance on the first card (780ti) which ends up with the gpu usage around 47% and does about 4.5MH
(instead of 11MH on the forthcoming ccminer...)

[go6ooo1212]

Im getting about 3050-3080 Kh/s per card with Palit 750TI stormX OC with the new miner. Im with 340.52 Nvidia driver, and this is my command line:
Cryptonite-CudaMiner-compute_35.exe -o stratum+tcp://xcnpool.1gh.com:7333 -u Y -p x -t 6
I've tried with compute_30 but the speed is about 2700kh.
How could someone reach 4MH ?

Actually I can, (my card is OCed), however the 750ti being a second card, it kills at the same time the performance on the first card (780ti) which ends up with the gpu usage around 47% and does about 4.5MH
(instead of 11MH on the forthcoming ccminer...)

Cant wait to try the new ccminer

[catia]

Im getting about 3050-3080 Kh/s per card with Palit 750TI stormX OC with the new miner. Im with 340.52 Nvidia driver, and this is my command line:
Cryptonite-CudaMiner-compute_35.exe -o stratum+tcp://xcnpool.1gh.com:7333 -u Y -p x -t 6
I've tried with compute_30 but the speed is about 2700kh.
How could someone reach 4MH ?

Actually I can, (my card is OCed), however the 750ti being a second card, it kills at the same time the performance on the first card (780ti) which ends up with the gpu usage around 47% and does about 4.5MH
(instead of 11MH on the forthcoming ccminer...)

I believe there is a bug in the MyStreamSynchronize usage that is causing the sleep time calculation to apply globally instead of per thread. So mixing cards of different performance causes problems ranging from high cpu utilization to under gpu utilization depending on how large the speed difference is between cards. Will look into it later.