896 mh/s firmware release - Butterfly Labs

[Buckwheet]

Flashed one cool one of mine and it is now 857 up from 816.  Temps still around 56 degrees.   Very nice work BFL.

Flashed one of mine and also getting 857, but my temps are 49C.

[xDGDZEx]

Flashed one cool one of mine and it is now 857 up from 816.  Temps still around 56 degrees.   Very nice work BFL.

Flashed one of mine and also getting 857, but my temps are 49C.

Third person also getting 857 MH. Out of 6 rev3's and 2 rev2's, three of the rev3's will run (at 857) without throttling.

Temps are ~55,57,62C on the ones with the higher clocks.

Also interesting note, all units are spaced roughly the same distance apart, and the 857 units do not have significantly higher temps than the stock units.

[Fefox]

so far 857.7Mh/s with the bitminter miner and no throttling....

when does the 900 Mh/s firmware come out? 

[nbtcminer]

Flashed one cool one of mine and it is now 857 up from 816.  Temps still around 56 degrees.   Very nice work BFL.

Flashed one of mine and also getting 857, but my temps are 49C.

Third person also getting 857 MH. Out of 6 rev3's and 2 rev2's, three of the rev3's will run (at 857) without throttling.

Temps are ~55,57,62C on the ones with the higher clocks.

Also interesting note, all units are spaced roughly the same distance apart, and the 857 units do not have significantly higher temps than the stock units.

In a fairly cold room; also getting 857.7 at around 43-43.7C. Nice work BFL!

[Epoch]

This may be surprising to some, but I'm generally more interested in the slower firmware.

I've replaced the stock fans with (much) quieter ones, accepting that some units will occasionally throttle due to the lowered CFM. By loading the 800 or 792 firmware onto those marginal units, I'm able to eliminate throttling and increase their effective speed.

The various firmware versions give me a way to optimize speed vs. noise that wasn't available before. Much appreciated.

[bombo999]

has anyone else encountered a firmware integrity failure ... after attempting to upgrade the firmware ... the following message is listed in the device information ... "Firmware integrity test failed.  Reflash required."

[BFL-Engineer]

has anyone else encountered a firmware integrity failure ... after attempting to upgrade the firmware ... the following message is listed in the device information ... "Firmware integrity test failed.  Reflash required."

This happens if during the firmware upload something goes wrong ( power failure, device disconnect, etc).
Once this happens, the unit must be re-flashed.


Regards,
BF Labs Inc.

[jddebug]

Although I'm seeing more MH/s wit the 863 firmware, My "U" in cgminer has gone down and the pools reported GH/s is down too. My average MH/s as reported by cgminer is up about 300MH/s across my 10 singles.

Any thoughts on why the cgminer reported MH/s is up but less work being reported by the pool?

[kano]

Although I'm seeing more MH/s wit the 863 firmware, My "U" in cgminer has gone down and the pools reported GH/s is down too. My average MH/s as reported by cgminer is up about 300MH/s across my 10 singles.

Any thoughts on why the cgminer reported MH/s is up but less work being reported by the pool?

Most likely: U is random - just like block finding.
Check it again in a few days

[jddebug]

Although I'm seeing more MH/s wit the 863 firmware, My "U" in cgminer has gone down and the pools reported GH/s is down too. My average MH/s as reported by cgminer is up about 300MH/s across my 10 singles.

Any thoughts on why the cgminer reported MH/s is up but less work being reported by the pool?

Most likely: U is random - just like block finding.
Check it again in a few days

Ok, but they have always reported as 8+GH/s on the pools but now only 7.x GH/s.

[-ck]

Although I'm seeing more MH/s wit the 863 firmware, My "U" in cgminer has gone down and the pools reported GH/s is down too. My average MH/s as reported by cgminer is up about 300MH/s across my 10 singles.

Any thoughts on why the cgminer reported MH/s is up but less work being reported by the pool?

Most likely: U is random - just like block finding.
Check it again in a few days

Ok, but they have always reported as 8+GH/s on the pools but now only 7.x GH/s.

The pools report based on your U. So if you're unlucky for a while, your U goes down, and your pool reports it slower. Doesn't mean you're actually hashing slower for that period.

[jddebug]

Although I'm seeing more MH/s wit the 863 firmware, My "U" in cgminer has gone down and the pools reported GH/s is down too. My average MH/s as reported by cgminer is up about 300MH/s across my 10 singles.

Any thoughts on why the cgminer reported MH/s is up but less work being reported by the pool?

Most likely: U is random - just like block finding.
Check it again in a few days

Ok, but they have always reported as 8+GH/s on the pools but now only 7.x GH/s.

The pools report based on your U. So if you're unlucky for a while, your U goes down, and your pool reports it slower. Doesn't mean you're actually hashing slower for that period.

Thank you for the explanation. Makes perfect sense. Does the U also go down when the difficulty goes up?

[kano]

Although I'm seeing more MH/s wit the 863 firmware, My "U" in cgminer has gone down and the pools reported GH/s is down too. My average MH/s as reported by cgminer is up about 300MH/s across my 10 singles.

Any thoughts on why the cgminer reported MH/s is up but less work being reported by the pool?

Most likely: U is random - just like block finding.
Check it again in a few days

Ok, but they have always reported as 8+GH/s on the pools but now only 7.x GH/s.

The pools report based on your U. So if you're unlucky for a while, your U goes down, and your pool reports it slower. Doesn't mean you're actually hashing slower for that period.

Thank you for the explanation. Makes perfect sense. Does the U also go down when the difficulty goes up?

Statistically - no.
(but of course, the value of your shares will drop)

[-ck]

Thank you for the explanation. Makes perfect sense. Does the U also go down when the difficulty goes up?

The utility is based on the difficulty shares you are mining for. Most pools work with difficulty one shares, regardless of the overall bitcoin block difficulty. About the only time you are mining for different difficulty shares is when you're mining on p2pool. (unrelated) However p2pool is a terrible choice to mine with these devices on because of the ~25% work lost.

[nbtcminer]

This may be surprising to some, but I'm generally more interested in the slower firmware.

I've replaced the stock fans with (much) quieter ones, accepting that some units will occasionally throttle due to the lowered CFM. By loading the 800 or 792 firmware onto those marginal units, I'm able to eliminate throttling and increase their effective speed.

The various firmware versions give me a way to optimize speed vs. noise that wasn't available before. Much appreciated.

@Epoch:

Just curious, what brand / model fan did you replace the stock one with?


-nbtcminer

[abbeytim]

heres screenshot with new firmware https://i.imgur.com/dqOYX.jpg 

[Epoch]

This may be surprising to some, but I'm generally more interested in the slower firmware.

I've replaced the stock fans with (much) quieter ones, accepting that some units will occasionally throttle due to the lowered CFM. By loading the 800 or 792 firmware onto those marginal units, I'm able to eliminate throttling and increase their effective speed.

The various firmware versions give me a way to optimize speed vs. noise that wasn't available before. Much appreciated.

Just curious, what brand / model fan did you replace the stock one with?

I've tried nearly 10 different fans on my Singles; the one I ended up using is the Silenx Effizio EFX-09-15. For what I was looking for, it had the best balance of silence and airflow of the models I've tried.

Out of nine 832Mhps Singles that I have so far retrofitted, 4 run without throttling even up to 27C ambient but the 5 others throttle.

For the 5 throttling ones, I've loaded 800Mhps firmware which allows them to run full-speed (no throttling). Your mileage may vary, of course.

[xDGDZEx]

This may be surprising to some, but I'm generally more interested in the slower firmware.

I've replaced the stock fans with (much) quieter ones, accepting that some units will occasionally throttle due to the lowered CFM. By loading the 800 or 792 firmware onto those marginal units, I'm able to eliminate throttling and increase their effective speed.

The various firmware versions give me a way to optimize speed vs. noise that wasn't available before. Much appreciated.

I too, am glad for this. Three of my singles would do this thing where they would clearly throttle (reduced output on cgminer) but they wouldn't blink their LEDs. I just installed the 816 MH firmware and now they are sitting at 811 MH and are rock solid. Actually increased my (avg) hash rate from ~6200 to ~6500

[jddebug]

heres screenshot with new firmware https://i.imgur.com/dqOYX.jpg 

 I posted this on another thread too but its best here.

864 runs great on mine too. However, even though the Hash rate is increased, the performance is decreased. I can not figure out why but when I went back to 832 I am doing better than at 862. If mine were throttling I could not tell. No blinking light and hash rate was at expected rate. Just very low U.

[Cablez]

If you run the easyminer test sequence it will tell you if the unit throttles.  Maybe that is worth a try.