Topic: FPGA development board "Icarus" - DisContinued/ important announcement - page 40. (Read 207279 times)
January 02, 2012, 08:39:10 AM
OK , all updated @ 1st floor.
January 01, 2012, 10:23:25 AM
sorry guys, today i took some holiday time. the order/ pre-order looks like will delay for some hours.
i received huge numbers of email and PM for order, but the order flow is really not started yet.
notice that i reserve the rights to choice my customers. so it's no need to feel worry about the pre-order.
i received huge numbers of email and PM for order, but the order flow is really not started yet.
notice that i reserve the rights to choice my customers. so it's no need to feel worry about the pre-order.
January 01, 2012, 12:29:29 AM
hi guys, here is 2 hours to the new year! happy new year everyone!
i just released a beta ver fpga code and miner software. speed up to 380Mh/s per board, and power is (looks like) equal as the old one. long pool is supported. and the logic is very robust.
the source will release soon.
10 boards is under test.
i just released a beta ver fpga code and miner software. speed up to 380Mh/s per board, and power is (looks like) equal as the old one. long pool is supported. and the logic is very robust.
the source will release soon.
10 boards is under test.
Awesome. I have no doubt that given enough time you will hit 200MH/s per chip. Was 80%, now it is 90%.
notice there is a -2 device on Icarus.... any body else is using a -3 device...
So are you saying it might be possible to push BEYOND 200 MH/s? If so then even more awesome.
oh ,the opposite.
-2 which i'm using is 15-20% SLOWER than -3 device.
but cheap.
I see. Even more impressive to get within 10% of ztex w/ a slower speed rated chip. Very impressive.
January 01, 2012, 12:25:53 AM
hi guys, here is 2 hours to the new year! happy new year everyone!
i just released a beta ver fpga code and miner software. speed up to 380Mh/s per board, and power is (looks like) equal as the old one. long pool is supported. and the logic is very robust.
the source will release soon.
10 boards is under test.
i just released a beta ver fpga code and miner software. speed up to 380Mh/s per board, and power is (looks like) equal as the old one. long pool is supported. and the logic is very robust.
the source will release soon.
10 boards is under test.
Awesome. I have no doubt that given enough time you will hit 200MH/s per chip. Was 80%, now it is 90%.
notice there is a -2 device on Icarus.... any body else is using a -3 device...
So are you saying it might be possible to push BEYOND 200 MH/s? If so then even more awesome.
oh ,the opposite.
-2 which i'm using is 15-20% SLOWER than -3 device.
but cheap.
ADD:
the small fan is not defaultly installed. this time.
January 01, 2012, 12:20:28 AM
Nice work on getting the hash rate up!
Are you referring to the number of FPGAs? I haven't seen anybody with 3FPGAs on a board. I think rph is working on a 4FPGA board, but haven't heard if it is working.
If you're referring to the grade of the FPGA, the X6500 has 2 -3, but as far as I know they are getting around ~280MH/S for the whole board. I thought you're using a -3?
Quote
notice there is a -2 device on Icarus.... any body else is using a -3 device...
Are you referring to the number of FPGAs? I haven't seen anybody with 3FPGAs on a board. I think rph is working on a 4FPGA board, but haven't heard if it is working.
If you're referring to the grade of the FPGA, the X6500 has 2 -3, but as far as I know they are getting around ~280MH/S for the whole board. I thought you're using a -3?
no, -2 and -3 refers to the device speed grade. -3 is 15-20% faster than -2 device.
December 31, 2011, 11:10:42 PM
Nice work on getting the hash rate up!
Are you referring to the number of FPGAs? I haven't seen anybody with 3FPGAs on a board. I think rph is working on a 4FPGA board, but haven't heard if it is working.
If you're referring to the grade of the FPGA, the X6500 has 2 -3, but as far as I know they are getting around ~280MH/S for the whole board. I thought you're using a -3?
Quote
notice there is a -2 device on Icarus.... any body else is using a -3 device...
Are you referring to the number of FPGAs? I haven't seen anybody with 3FPGAs on a board. I think rph is working on a 4FPGA board, but haven't heard if it is working.
If you're referring to the grade of the FPGA, the X6500 has 2 -3, but as far as I know they are getting around ~280MH/S for the whole board. I thought you're using a -3?
December 31, 2011, 11:06:46 PM
hi guys, here is 2 hours to the new year! happy new year everyone!
i just released a beta ver fpga code and miner software. speed up to 380Mh/s per board, and power is (looks like) equal as the old one. long pool is supported. and the logic is very robust.
the source will release soon.
10 boards is under test.
i just released a beta ver fpga code and miner software. speed up to 380Mh/s per board, and power is (looks like) equal as the old one. long pool is supported. and the logic is very robust.
the source will release soon.
10 boards is under test.
Awesome. I have no doubt that given enough time you will hit 200MH/s per chip. Was 80%, now it is 90%.
notice there is a -2 device on Icarus.... any body else is using a -3 device...
So are you saying it might be possible to push BEYOND 200 MH/s? If so then even more awesome.
December 31, 2011, 10:57:30 PM
hi guys, here is 2 hours to the new year! happy new year everyone!
i just released a beta ver fpga code and miner software. speed up to 380Mh/s per board, and power is (looks like) equal as the old one. long pool is supported. and the logic is very robust.
the source will release soon.
10 boards is under test.
i just released a beta ver fpga code and miner software. speed up to 380Mh/s per board, and power is (looks like) equal as the old one. long pool is supported. and the logic is very robust.
the source will release soon.
10 boards is under test.
Awesome. I have no doubt that given enough time you will hit 200MH/s per chip. Was 80%, now it is 90%.
notice there is a -2 device on Icarus.... any body else is using a -3 device...
December 31, 2011, 05:22:48 PM
hi guys, here is 2 hours to the new year! happy new year everyone!
i just released a beta ver fpga code and miner software. speed up to 380Mh/s per board, and power is (looks like) equal as the old one. long pool is supported. and the logic is very robust.
the source will release soon.
10 boards is under test.
i just released a beta ver fpga code and miner software. speed up to 380Mh/s per board, and power is (looks like) equal as the old one. long pool is supported. and the logic is very robust.
the source will release soon.
10 boards is under test.
Awesome. I have no doubt that given enough time you will hit 200MH/s per chip. Was 80%, now it is 90%.
December 31, 2011, 10:42:08 AM
Happy 2012 everyone. This year will bring some greate advancement with fpga's.
December 31, 2011, 08:36:15 AM
Happy New Year everybody. May 2012 be a good year for you!
December 31, 2011, 08:28:50 AM
hi guys, here is 2 hours to the new year! happy new year everyone!
i just released a beta ver fpga code and miner software. speed up to 380Mh/s per board, and power is (looks like) equal as the old one. long pool is supported. and the logic is very robust.
the source will release soon.
10 boards is under test.
Happy new year! i just released a beta ver fpga code and miner software. speed up to 380Mh/s per board, and power is (looks like) equal as the old one. long pool is supported. and the logic is very robust.
the source will release soon.
10 boards is under test.
December 31, 2011, 08:27:03 AM
also a good new year to you!
2012 will offer many many new suprises for us!
best regards from one in the year 2011...
2012 will offer many many new suprises for us!
best regards from one in the year 2011...
December 31, 2011, 08:14:31 AM
hi guys, here is 2 hours to the new year! happy new year everyone!
i just released a beta ver fpga code and miner software. speed up to 380Mh/s per board, and power is (looks like) equal as the old one. long pool is supported. and the logic is very robust.
the source will release soon.
10 boards is under test.
i just released a beta ver fpga code and miner software. speed up to 380Mh/s per board, and power is (looks like) equal as the old one. long pool is supported. and the logic is very robust.
the source will release soon.
10 boards is under test.
December 29, 2011, 10:57:08 AM
Quote
What about an upgraded chip? Is it worth it to go to something with more gates?
I browsed through digi-key and it seems like anything with more gates costs in the thousands per chip. My guess is that the spartan 6 offers the best bang for the buck ATM since it is kind of old but I could be wrong.
Quote
warranty is invalid for this product. before shipping, all boards test for weeks. and please test them in 48Hrs after you received them. if any hardware fault happened, you can send it to me and i'll repair it if i can, you only need pay the parts cost and factory cost and shipment cost.
My understanding is that the boards are tested before shipping but there are no warranties offered. https://bitcointalksearch.org/topic/fpga-development-board-icarus-discontinued-important-announcement-51371
The LX150 is the largest part available in the Spartan-6 Family. For the time being, the Spartan-6 is probably the most cost effective FPGA for bitcoin mining. The Altera Cyclone IV might be close, but it hasn't had the development effort poured into it like the spartan has.
Unless someone comes up with a (much) more efficient sha-256 algorithm, the Spartan-6 will remain the most cost efficient part until the next generation of FPGAs are (widely) available. These include parts like the Xilinx Artix. I don't expect the Artix to be available in production quantities for at least another 6 months, and more likely, a year or more.
For the time being, it's going to be difficult to beat the Spartan - although I think the Cyclone IV deserves some additional attention. Even if the Cyclone can beat out the Spartan though - it's not going to be by any remarkable amount. The next-gen parts, however, should be able to manage a pretty good margin over the current Spartan/Cyclone parts.
Enigma
I was looking at the Artix-7. It's twice the number of gates form the spartan-6. right now it looks like you could fit 2.5 almost 3 hashing branches into the147k gates so with 330k gates I would hope you can get 6.
-IT
Sure except you aren't getting any Artix-7 at any price in volume unless your sister is married to the CEO of xilinix or your company does a couple billion in sales with them.
That is what people mean by 'next gen'. Artix-7 is a 28nm chip. It is available in low volume for testing & analsyis by prefered partners. It likely won't be available in volume until Q2 2012. The smaller you are the longer you likely will wait and the more of a premium you will pay. Even then there is no guarantee the Artix-7 would be cheaper. It would need to be priced below $500 to be more cost effective than Spartan-6. Likely it is be >$1000 for a some time. Maybe even into 2013.
For the time being there is no chip AVAILABLE which is more cost effective than the Spartan 6 and the LX150 is the largest one in the line.
December 29, 2011, 10:20:47 AM
Quote
For the time being, it's going to be difficult to beat the Spartan - although I think the Cyclone IV deserves some additional attention. Even if the Cyclone can beat out the Spartan though - it's not going to be by any remarkable amount. The next-gen parts, however, should be able to manage a pretty good margin over the current Spartan/Cyclone parts.
Good point on the Cyclone. I think fpgaminer compared the 2. Has anybody seen any non-eval boards using the Cyclone?
Quote
I was looking at the Artix-7. It's twice the number of gates form the spartan-6. right now it looks like you could fit 2.5 almost 3 hashing branches into the147k gates so with 330k gates I would hope you can get 6.
It would be interesting to see how the new generation of Xilinx 7's do in hashing. My understanding is that the Artix and the Virtex series are not not widely available yet. 6 months to a year would be my guess too. It looks like they are/will be fab by TSMC on the 28nm so that should start to ramp next year. The Kintex7's with 326K gates are selling for $1560/chip with limited quantity so let's hope the Artix can deliver on price.
December 29, 2011, 07:50:22 AM
Quote
What about an upgraded chip? Is it worth it to go to something with more gates?
I browsed through digi-key and it seems like anything with more gates costs in the thousands per chip. My guess is that the spartan 6 offers the best bang for the buck ATM since it is kind of old but I could be wrong.
Quote
warranty is invalid for this product. before shipping, all boards test for weeks. and please test them in 48Hrs after you received them. if any hardware fault happened, you can send it to me and i'll repair it if i can, you only need pay the parts cost and factory cost and shipment cost.
My understanding is that the boards are tested before shipping but there are no warranties offered. https://bitcointalksearch.org/topic/fpga-development-board-icarus-discontinued-important-announcement-51371
The LX150 is the largest part available in the Spartan-6 Family. For the time being, the Spartan-6 is probably the most cost effective FPGA for bitcoin mining. The Altera Cyclone IV might be close, but it hasn't had the development effort poured into it like the spartan has.
Unless someone comes up with a (much) more efficient sha-256 algorithm, the Spartan-6 will remain the most cost efficient part until the next generation of FPGAs are (widely) available. These include parts like the Xilinx Artix. I don't expect the Artix to be available in production quantities for at least another 6 months, and more likely, a year or more.
For the time being, it's going to be difficult to beat the Spartan - although I think the Cyclone IV deserves some additional attention. Even if the Cyclone can beat out the Spartan though - it's not going to be by any remarkable amount. The next-gen parts, however, should be able to manage a pretty good margin over the current Spartan/Cyclone parts.
Enigma
I was looking at the Artix-7. It's twice the number of gates form the spartan-6. right now it looks like you could fit 2.5 almost 3 hashing branches into the147k gates so with 330k gates I would hope you can get 6.
-IT
December 29, 2011, 06:07:50 AM
Quote
What about an upgraded chip? Is it worth it to go to something with more gates?
I browsed through digi-key and it seems like anything with more gates costs in the thousands per chip. My guess is that the spartan 6 offers the best bang for the buck ATM since it is kind of old but I could be wrong.
Quote
warranty is invalid for this product. before shipping, all boards test for weeks. and please test them in 48Hrs after you received them. if any hardware fault happened, you can send it to me and i'll repair it if i can, you only need pay the parts cost and factory cost and shipment cost.
My understanding is that the boards are tested before shipping but there are no warranties offered. https://bitcointalksearch.org/topic/fpga-development-board-icarus-discontinued-important-announcement-51371
The LX150 is the largest part available in the Spartan-6 Family. For the time being, the Spartan-6 is probably the most cost effective FPGA for bitcoin mining. The Altera Cyclone IV might be close, but it hasn't had the development effort poured into it like the spartan has.
Unless someone comes up with a (much) more efficient sha-256 algorithm, the Spartan-6 will remain the most cost efficient part until the next generation of FPGAs are (widely) available. These include parts like the Xilinx Artix. I don't expect the Artix to be available in production quantities for at least another 6 months, and more likely, a year or more.
For the time being, it's going to be difficult to beat the Spartan - although I think the Cyclone IV deserves some additional attention. Even if the Cyclone can beat out the Spartan though - it's not going to be by any remarkable amount. The next-gen parts, however, should be able to manage a pretty good margin over the current Spartan/Cyclone parts.
Enigma
December 29, 2011, 01:29:54 AM
Quote
What about an upgraded chip? Is it worth it to go to something with more gates?
I browsed through digi-key and it seems like anything with more gates costs in the thousands per chip. My guess is that the spartan 6 offers the best bang for the buck ATM since it is kind of old but I could be wrong.
Quote
warranty is invalid for this product. before shipping, all boards test for weeks. and please test them in 48Hrs after you received them. if any hardware fault happened, you can send it to me and i'll repair it if i can, you only need pay the parts cost and factory cost and shipment cost.
My understanding is that the boards are tested before shipping but there are no warranties offered. https://bitcointalksearch.org/topic/fpga-development-board-icarus-discontinued-important-announcement-51371
December 29, 2011, 12:13:53 AM
Quote from: ngzhang
pre-order/order will atart @ 1/1/2012
Sorry if this has been mentioned already somewhere, but how can we ensure we get in the next pre-order? E-mail you or PM? Do you have a btc address to send the first deposit to, etc?
Jump to: