Topic: Discontinued: Lancelot (Heavy Duty 400 Mega hashes, 2 x FPGA Mining Board) - page 14. (Read 86497 times)

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.

thanks for explanation Kano. reconfigured.)

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:
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

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

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

yes. definitely..

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

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

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

What quad chip boards?

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

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

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?

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

Blackarrow, i made order #1771 and want to know how can we use bitcoin escrow to pay for it ?

Thks

No, of course, I know)) My PSU has 4 modular generic connectors and I have 4 leads x4 Molex connectors on the ends.

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

Thank you for letting me know!

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

Can somebody confirm that Molex connector is working, so I can connect the board to a regular PSU? Thanks