Discontinued: Lancelot (Heavy Duty 400 Mega hashes, 2 x FPGA Mining Board)

[wedgy2k]

Hi

Will the Lancelot (Group Manufacturing Shipping 15th July) be available as a collection from Reading UK?
Could the remaining 50% be payed on collection?
You have a good rep on here and I'd rather you hold 50% of my order money than some of the "other" company's selling mining kit!

Cheers

[Pistachio]

Yes, the 4 pin generic connector works just fine. You can daisy chain them for multiple cards.

Don't daisy chain the power cables unless you really want to burn down your house.

Good to know. I have a two way splitter on two different molex 4pin connectors from PSU to power 4 boards. The other 10 are powered through the ATX adapter. Is this a potential fire hazard?

[setkeh]

Yes, the 4 pin generic connector works just fine. You can daisy chain them for multiple cards.

Don't daisy chain the power cables unless you really want to burn down your house.

Good to know. I have a two way splitter on two different molex 4pin connectors from PSU to power 4 boards. The other 10 are powered through the ATX adapter. Is this a potential fire hazard?

Hmm its quite possible though you should feel the cables themselves and also the exhaust from the PSU fans to see if either are Really hot.
If its hot to the touch then on average its too hot (most people can stand temps between 60-90C on there hardened fingertips

[Bogart]

4 dual-chip boards off of one 18-ga wire (pair) is OK.  4 quad-chip boards isn't.

[kano]

4 dual-chip boards off of one 18-ga wire (pair) is OK.  4 quad-chip boards isn't.

What quad chip boards?

[bitpop]

It's where you glue two lancelot together :-P

[jimmy3dita]

Hello everyone, I wish to share my quick feedback about the FPGAs and the new V42 firmware.

My configuration (following the precious feedback from Kano) is --icarus-timing 2.0=75 and I get way better results than simply chosing "short" or "long".

Note that the ICA 0 has the V42 firmware (it will last there for 24 hours, then I'll decide if apply the upgrade to all other FPGAs) and the difficulty is 2:



Any hints are welcome, I'd like to calculate the hashrate of the V42 one

[bitpop]

Is.--icarus-timing 2.0=75 better than short for these units?

[Xialla]

Is.--icarus-timing 2.0=75 better than short for these units?

yes. definitely..

[Schrankwand]

Hello everyone, I wish to share my quick feedback about the FPGAs and the new V42 firmware.

My configuration (following the precious feedback from Kano) is --icarus-timing 2.0=75 and I get way better results than simply chosing "short" or "long".

Note that the ICA 0 has the V42 firmware (it will last there for 24 hours, then I'll decide if apply the upgrade to all other FPGAs) and the difficulty is 2:



Any hints are welcome, I'd like to calculate the hashrate of the V42 one

You have the U Rating there. 2.89 seems to work as an average one. 3.69/2.89= 127,7%.

--> Nehmen wir die 420 als Basis, oder 399, wie angepeilt, kommen wir auf --> 536 mhash/s bzw 509mh/s

[bitpop]

I've seen so many, 2.5=100 etc does it really matter

Is.--icarus-timing 2.0=75 better than short for these units?

yes. definitely..

[kano]

Is.--icarus-timing 2.0=75 better than short for these units?

yes. definitely..

IF you are referring to the standard bitstream ...

No.

Firstly: it should be --icarus-timing 2.5=100 (not --icarus-timing 2.0=100)

Since it runs at 400MH/s setting it to 2.0 will make the MH/s read higher than it really is.

Secondly: 100 vs 75 will not make much of a difference ... but the minor difference is that 100 is better than 75
75 vs 100: 75 means that it will abort work at 7.5 seconds and start a new work item
100 means 10.0 seconds
Thus 75 will use 33% more work items for the same amount of work and thus 33% more CPU for work generation in cgminer.

bitpop, the easiest way to see this is to run it --icarus-timing short and look at the "Re-estimate:"s it produces.

For my one, a complete short run gives:

Code:

 [2013-05-27 22:30:04] BLT0 Re-estimate: Hs=2.500162e-09 W=6.150422e-03 read_time=10594ms fullnonce=10.744s
 [2013-05-27 22:31:32] BLT0 Re-estimate: Hs=2.500081e-09 W=6.179870e-03 read_time=10593ms fullnonce=10.744s
 [2013-05-27 22:36:02] BLT0 Re-estimate: Hs=2.500034e-09 W=6.350961e-03 read_time=10593ms fullnonce=10.744s
 [2013-05-27 22:41:43] BLT0 Re-estimate: Hs=2.500021e-09 W=6.256367e-03 read_time=10593ms fullnonce=10.744s
 [2013-05-27 22:57:04] BLT0 Re-estimate: Hs=2.500064e-09 W=6.206789e-03 read_time=10593ms fullnonce=10.744s
 [2013-05-27 23:23:27] BLT0 Re-estimate: Hs=2.500060e-09 W=6.225088e-03 read_time=10593ms fullnonce=10.744s

So you can clearly see it is 2.5 and 10s is under the 10.744s by 3/4 of a second so it's fine.

Note that above is the new code.
... which I've finally fixed all the windows MTX locking and hotplug problems that I know about that seem to have been hiding from me for a while ...

...

Now for  jimmy3dita's results:
If he is using a ~500MH/s bitstream then he should be using: --icarus-timing 2.0=75
As I've said before the 75 number has to be less than the fastest speed it will run at.
For 500MH/s that's 8.5899s
However if it is running a bit faster it will get idle if the number used is more than the hashing time.
75 means it will be OK for up to 572MH/s so it gives some leeway for the bitstream to adjust the clocking faster but still never fall idle.

Hmm I seem to be repeating this

[Xialla]

thanks for explanation Kano. reconfigured.)

[bitpop]

Thank you amazing kano. I did see higher mhs but it was probably fake. I still don't know what I'm doing
 so I'm sticking with short. Yes I'm using standard rom and p2pool, I need it to abort quickly.

[bitpop]

Can I just keep using short and have it do it's thing?

[kano]

Can I just keep using short and have it do it's thing?

Yes.

The only advantage of not using short is that if there is some major computer performance hiccup during the ~1hr that short takes to run, it may get something strange as the performance figures it uses from 1hr forever after until you restart it.

[kano]

Thank you amazing kano. I did see higher mhs but it was probably fake. I still don't know what I'm doing
 so I'm sticking with short. Yes I'm using standard rom and p2pool, I need it to abort quickly.

Forcing it to abort earlier doesn't make any advantage for p2pool.
It checks every 100ms for an LP and aborts the work early if an LP occurs.

[bitpop]

Last question. Short vs long?

[jimmy3dita]

For sure "short timing" is the best choice on the older firmware, with the new one I see orange leds (idle) randomly blinking, meaning that it's possible to squeeze a bit more from the chips

[squall1066]

Can anyone here help me out on this please.

https://bitcointalk.org/index.php?topic=216792.0

I got help from blackarrow and he sorted in 5 min, I have had a drive crash, Now I cant get it going again, As I have no clue what he did.