My guess is that the calculations are hard enough, and L2 catch but enough in modern CPUs that RAM access may not be a huge bottleneck.
Just guessing thoigh.
My guess is that the calculations are hard enough, and L2 catch but enough in modern CPUs that RAM access may not be a huge bottleneck.
Just guessing thoigh.
At least for 4 cores I have not been able to see a slow up due to the RAM bandwidth over using 1 core. It might be that up to 8 cores will be beneficial. It depends on how fast your RAM and CPU are. For a typical machine, half the time is spent on the CPU and the other half accessing RAM. CPU cache must be helping the cores to prevent colliding when accessing RAM. You can see it in all the benchmarks people have reported since at least z7. A 2 GHz CPU with DDR4 is not as fast on the benchmark as a 3 GHz CPU with DDR3.
blocks gains = CPU GHz x RAM B.W. x cores
The calculations are modest (not easy, as choice of CPU does make a difference) but the dataset does not have to be much larger than the CPU caches to cause very frequent cache misses and force constant system memory access.
It takes very few threads to overwhelm available memory bandwidth.
A GPU programmer only needs to state his conditions, then let Austin-Williams be the escrow. Or maybe I would buy one for $10k, then sell it one at a time.
Do we want to simply approach Tim and the cloud mining gang with an offer and buy their GPU miner?
The DAO 2.0? No, thank. Don’t try fooling me // I’m not troll
Also, you can/may try talk with in # zcash slack.com chanal (if you can)
Hi everyone, any news about this GPU mining software? Will it be available in October?
The developer sold to a private company. No new developer has come forward.
I would like to contribute 1 BTC for GPU miner…
but how can I join in the crowdfund?
You just did. We’re just waiting on a developer, any developer, to state they are willing to develop one either for us as a group or for open source.
@zawy thank you for your answer.
Now the point is that we need one.
What if we post this thread to some good miner developers on bitcointalk and see what they say?
Have at it. It’s in my best interest if there are no GPU miners, especially an open one, because I’m already invested in CPUs that for the most part can’t hold GPUs. I don’t suppose there’s a decent USB GPU (ha ha?).
I see, well something like this should do the job I guess Crank Up Your Laptop's Gaming Power with an External Video Card Dock
Thanks, but judging by the other GPU miner thread with the first public info on a GPU miner, it looks like it’s going to be cheaper to stick with PCs unless you already have GPU’s laying around.
This is really bad!!!
Actually, I’d say we are waiting for money to actually be in escrow. Also we need a clear & finite list of features needed to have the funds released to a dev.
That’s what is needed to get interested developers. I’ve talked to two developers – one with over 2 decades of parallel programing experience (was one of the first people building FPGAs to break DES) – and they basically told me “tell me when there is money in escrow” and I’ll take a look.
I suspect most other devs feel the same way.
Who is the escrow? Do we have anybody trustworthy? We have enough people ready to donate the money, we just need an escrow.
I have contacted somebody and he will be doing both CPU and GPU software this week. If his results are better than what is available we have our guy.
I’d be happy to put my 1 BTC in escrow if you can set one up.