Topic: [XPM] mtrlt Primecoin OCL GPU Miner, Communication problem ??? (Read 1444 times)
Geefe-a me-a muney und I veell creete-a yuoo a beooteeffool grepheecs meener fur preemecuin. Zeen yuoo veell ell injuy it es I hefe-a beee fur a foo deys elreedy oor I veell deeseppeer veet yuoor duneshuns Cheesy BORK BORK BORK!
In the beginning of Prime searching, CUDA was first tapped in Prime95, And though it is multistreamed it didn't look any faster than any conventional CPU. Right now, It will take a huge process and a little setback on the multiple process in generating prime numbers whereas OpenCL is designed to make FLOPS faster, now the problem here is the process is not one pass, it is multiple pass, I wonder how the vCard can be 100% utilized when there is a big hiccup in the system? 
Well, you didn't PAY for anything. Donations are donations. Doesn't mean guarantee of goods or service. Also, nobody claimed GPU mining would be viable.
I was almost convinced to donate to this effort but luckily (for me anyhow) things started to look bad soon enough to ward me off. I thought this would happen for quiet some time. I wonder if it was ever the same person from the other miner or if it was just a scam from the beginning. There has been plenty of discussion into whether he actually owes anyone anything and he doesnt (at least in my mind) due to the fact that all payments were donations but if he ever was really working on this he would have stated what the problem was. Even if he gave up I'd think he would at least say something about this. He's just gone as it seems.
bump
It looks like an artificial delay release GPU Miner 
I've no stake in the game but I think when you're months late communicating efficiently should be the least you can do.
mtrlt has blocked his two thread to comments and the last update was :
https://bitcointalksearch.org/topic/m.3008508
Correct me if I'm wrong but how is a 100% increase on a GPU (HD6990) that is 400% more expensive than a Phenom x6 CPU an improvement in the slightest ?
Unoptimized preliminary results,, But still... any real world expectation at this time ?
mtrlt has blocked his two thread to comments and the last update was :
Quote
New update!!!
The GPU primality tester works now! Finally.. Testing the first, very much unoptimized version on testnet, a HD6990 seems to be about 2x as fast as a Phenom x6 1055T. On mainnet, there's still the problem of having too few candidates left after sieving, which leads to GPU underutilization.
I'll write a better update in the morning, it's 4 AM..
(4 days ago)The GPU primality tester works now! Finally.. Testing the first, very much unoptimized version on testnet, a HD6990 seems to be about 2x as fast as a Phenom x6 1055T. On mainnet, there's still the problem of having too few candidates left after sieving, which leads to GPU underutilization.
I'll write a better update in the morning, it's 4 AM..
https://bitcointalksearch.org/topic/m.3008508
Correct me if I'm wrong but how is a 100% increase on a GPU (HD6990) that is 400% more expensive than a Phenom x6 CPU an improvement in the slightest ?
Unoptimized preliminary results,, But still... any real world expectation at this time ?
Jump to: