Topic: FPGA worth buying? (Read 1264 times)
I want a LTC asic!!
Well, the real question is: Will you recover your cost before you can get the ASICs delivered?
Yes, probably, with the current valuation of Bitcoins you will, but who knows what tomorrow brings...
Yes, probably, with the current valuation of Bitcoins you will, but who knows what tomorrow brings...
i dont think FPGA will be reliable after the first big wave of asics was shipped. You should have started using FPGA one year ago, its def too late.
FPGA are re-programable, so might be worth something even if boom is gone.
Well, the real question is: Will you recover your cost before you can get the ASICs delivered?
Yes, probably, with the current valuation of Bitcoins you will, but who knows what tomorrow brings...
Yes, probably, with the current valuation of Bitcoins you will, but who knows what tomorrow brings...
there is ASIC DIY kit in develop now. may be u can you look into that
https://bitcointalk.org/index.php?topic=179769.180
but wont be so soon, 1st board only come out at around july
https://bitcointalk.org/index.php?topic=179769.180
but wont be so soon, 1st board only come out at around july
https://forum.litecoin.net/index.php/topic,2702.0/topicseen.html
Thats the official thread I will post info to when we get updates.
Thats the official thread I will post info to when we get updates.
There should be some juicy developments for the fpgas in the next couple weeks.
I'll be interested to see it. I've been looking into creating an FPGA-based miner and I'm not coming remotely close to the $/kh/s rates of ATI video cards. The salsa 20 core routine wants a lot of high-speed memory, and I just don't see how to get that with an FPGA. If you rely on the gate-array itself, you can get the raw speed you need but you don't have enough concurrent hashes and if you try to do something like use the FPGA DDR routines to go to a DDR, it's orders of magnitude too slow from a memory throughput perspective. It's the same fundamental issue that I think all of the FPGA investigations run into.
I will be extremely interested to see your approach because right now, with a ~$300 FPGA, I'm still ~1/3 of the dollar-to-hash rate that I'd need to be to compete with a GPU... although power looks better.
It's a gamble at best. Asics are only going to become more and more commonplace.
The fun thing about Field Programmable Gate Arrays is that they're field programmable - if you're at all so inclined, it's always fun to repurpose them for something new, like brute forcing password hashes.
I suppose.. but not for the prices they're selling for now.
You can better wait for and get half sized avalon.
The fun thing about Field Programmable Gate Arrays is that they're field programmable - if you're at all so inclined, it's always fun to repurpose them for something new, like brute forcing password hashes.
There should be some juicy developments for the fpgas in the next couple weeks.
Thanks for the info guys. Some of the old butterfly labs FPGAs seem to go for a lot still.
Yeah a couple on ebay past few days, I think ~$1300 which is way too rich for my blood.
Some sold months ago for a lot less, but that was before the BTC price surge.
the ltcfpga sites did not announce anything for months afaic
Thanks for the info guys. Some of the old butterfly labs FPGAs seem to go for a lot still.
Do you have any links to jasinlee's project? I've worked on FPGA's before and I'm curious about his approach.
https://forum.litecoin.net/index.php/topic,2702.0.html
Is most official thread to follow.
Do you have any links to jasinlee's project? I've worked on FPGA's before and I'm curious about his approach.
Will have to wait and see. There are 3 projects ongoing I believe, the main one I'm interested in is jasinlee's. His price goal is to match the $/hashrate of GPUs. I believe he has engineering samples delivered and is the furthest along but isn't providing specs, prices, preorders, etc until testing is finished and they are ready to produce a final product.
An LTC-based FPGA would be tricky and expensive and for equivalent hashing rates, it would be more expensive than a video card.
I'm quite interested in the multiple LTC FPGA projects though. Since ASICs don't exist for LTC yet, FPGAs can still make a lot of sense.
Jump to: