Avalon ASIC users thread - page 126.

SolarSilver

legendary

Activity: 1112

Merit: 1000

Quote from: craslovell on July 09, 2013, 07:13:20 PM

There's a ton of this in the logs

Jul 9 21:45:01 OpenWrt cron.info crond[885]: crond: USER root pid 13416 cmd /usr/bin/cgminer-monitor

Literally 100 of these messages in a row

That is normaal, it's a script run every 2 minutes (check with crontab -l from the root promt) that checks is your cgminer process is active, if not it restarts it

Code:

root@avalon00:~# crontab -l
*/2 * * * * /usr/bin/cgminer-monitor
root@avalon00:~#

craslovell

legendary

Activity: 1470

Merit: 1021

There's a ton of this in the logs

Jul 9 21:45:01 OpenWrt cron.info crond[885]: crond: USER root pid 13416 cmd /usr/bin/cgminer-monitor

Literally 100 of these messages in a row

craslovell

legendary

Activity: 1470

Merit: 1021

It seems to be stumbling here the past hour. Another 2 mins of no shares submitted... Connection appears fine. I am using wireless with a signal strength of 57% but I don't see any drops.

When it stops submitting shares the NMW value climbs quickly

craslovell

legendary

Activity: 1470

Merit: 1021

Quote from: Elokane on July 09, 2013, 06:56:30 PM

Quote from: craslovell on July 09, 2013, 06:50:14 PM

I know this was probably answered 50 times already in this thread and I tried looking, but it's a little bit of a TL;DR situation.

I have accepted 10124 and HW 202

From what I read the hardware errors are nothing to worry about, can anyone clarify this?

Also, I noticed NMW is now going up, there was a point where it didn't submit a share in about 2 minutes. My other miners are fine, only the avalon did this.

One more thing, since I am a noob to these, is it normal for the fans to randomly spin down a bit for a second or two and then go back to normal speed?

Nothing to worry about. You want to maximize speed so that the last step you take is the one before the step where you get more errors than accepted shares for.

Fans adjust based on temperature and so that's normal behaviour.

Thanks for the feedback, Temps are 23, 42, and 46. Highest fan speeds are 2600RPM. When the fans slow down it is literally for two second. They also go full blast for about two second periods. Is this strange behavior?

Hash rate jumps around a lot too, from what I've seen it is sometimes down at 35GH/s then up to 75GH/s.

20130419 is the firmware version and I am running "Extreme" at 300MHz

Elokane

hero member

Activity: 817

Merit: 1000

Truth is a consensus among neurons www.synereo.com

Quote from: craslovell on July 09, 2013, 06:50:14 PM

I know this was probably answered 50 times already in this thread and I tried looking, but it's a little bit of a TL;DR situation.

I have accepted 10124 and HW 202

From what I read the hardware errors are nothing to worry about, can anyone clarify this?

Also, I noticed NMW is now going up, there was a point where it didn't submit a share in about 2 minutes. My other miners are fine, only the avalon did this.

One more thing, since I am a noob to these, is it normal for the fans to randomly spin down a bit for a second or two and then go back to normal speed?

Nothing to worry about. You want to maximize speed so that the last step you take is the one before the step where you get more errors than accepted shares for.

Fans adjust based on temperature and so that's normal behaviour.

craslovell

legendary

Activity: 1470

Merit: 1021

I know this was probably answered 50 times already in this thread and I tried looking, but it's a little bit of a TL;DR situation.

I have accepted 10124 and HW 202

From what I read the hardware errors are nothing to worry about, can anyone clarify this?

Also, I noticed NMW is now going up, there was a point where it didn't submit a share in about 2 minutes. My other miners are fine, only the avalon did this.

One more thing, since I am a noob to these, is it normal for the fans to randomly spin down a bit for a second or two and then go back to normal speed?

mgio

hero member

Activity: 546

Merit: 500

Quote from: GenTarkin on July 09, 2013, 05:20:10 PM

lol, these things are nowhere near worth 140btc anymore but...okies =)

140btc might not be a horrible price, but I wouldn't pay much more than that. You might be able to ROI on 140 by the end of the year, especially if BFL continues to stall on their delivery.

GenTarkin

legendary

Activity: 2450

Merit: 1002

lol, these things are nowhere near worth 140btc anymore but...okies =)

ProfMac

legendary

Activity: 1246

Merit: 1002

Quote from: ?? on ??

Your second and third pool are still setup to mine for Avalon.
If 50btc goes down, you will automatically start mining for Avalon...
You should think about changing that, unless you really really really loved waiting for your machine/s.

But it leads to so much delicious misinterpretation that Avalon is secretly mining with unshipped units.

mgio

hero member

Activity: 546

Merit: 500

Quote from: ?? on ??

Your second and third pool are still setup to mine for Avalon.
If 50btc goes down, you will automatically start mining for Avalon...
You should think about changing that, unless you really really really loved waiting for your machine/s.

Quote

Yeah, I know. It's embarassing

. I actually changed it but I didn't "apply" it yet because I didn't want to restart the miner at that point because I was in the middle of testing power consumption with the kill-a-watt. I'll go reset that when I get home.

GenTarkin

legendary

Activity: 2450

Merit: 1002

Quote from: -ck on July 09, 2013, 04:12:57 PM

Quote from: GenTarkin on July 09, 2013, 04:10:28 PM

In case ya missed my edit:
In other words could one calculate the ratio of NMW to HW errors by doing (NMW / (HW errors - NMW)) * 100?

or would it be (NMW / HW errors) * 100

Either way, since they both are diff1 related work then either of the above should be valid .. right?

It will be (NMW / HW errors) * 100

Don't concern yourself too much by the distinction between the different type of hardware errors. Sometimes we just expose too much internal information that doesn't help the users much.

Yeah, I realize in the end all that matters is the cumulative amount between the 2 types, but I like the knowledge of knowing whats really goin on behind the scenes, so thanks for explaining it to me =)

-ck

legendary

Activity: 4088

Merit: 1631

Ruu \o/

Quote from: GenTarkin on July 09, 2013, 04:10:28 PM

In case ya missed my edit:
In other words could one calculate the ratio of NMW to HW errors by doing (NMW / (HW errors - NMW)) * 100?

or would it be (NMW / HW errors) * 100

Either way, since they both are diff1 related work then either of the above should be valid .. right?

It will be (NMW / HW errors) * 100

Don't concern yourself too much by the distinction between the different type of hardware errors. Sometimes we just expose too much internal information that doesn't help the users much.

GenTarkin

legendary

Activity: 2450

Merit: 1002

In case ya missed my edit:
In other words could one calculate the ratio of NMW to HW errors by doing (NMW / (HW errors - NMW)) * 100?

or would it be (NMW / HW errors) * 100

Either way, since they both are diff1 related work then either of the above should be valid .. right?

-ck

legendary

Activity: 4088

Merit: 1631

Ruu \o/

Quote from: GenTarkin on July 09, 2013, 04:02:53 PM

Ah, ok thanks for the explanation. In short OC'n can cause both types.
So would each NMW constitute as an invalid diff1 share? cuz afaik HW errors are all equivilant to invalid diff1 shares .. which is why it used as 2% of diff1shares for --avalon-auto...?

Yes each should have actually been a valid diff1share.

The 2% value is one I ~~pulled out of my arse~~ derived from a complicated regression analysis that correlated with the highest effective hashrate where pushing the MHz higher increased hardware errors and started decreasing valid share generation.

GenTarkin

legendary

Activity: 2450

Merit: 1002

Quote from: -ck on July 09, 2013, 03:58:51 PM

Quote from: GenTarkin on July 09, 2013, 03:56:16 PM

Quote from: -ck on July 09, 2013, 01:38:18 PM

Quote from: GenTarkin on July 09, 2013, 10:50:09 AM

Quote from: kano on July 09, 2013, 06:55:09 AM

I looked up the code (in relation to my HW error comments before)
HW is the count of bad nonces generated plus NMW
NMW is when work comes back but it doesn't match any work sent to it (thus cannot be used)

Anyway to minimize NMW?
My NMW hovers right around .52% of all matching work...

Or is NMW ALSO the result of chip errors?

Chip errors, even worse than regular HW errors.

Oh, so let me see if I understand, NMW is just chip errors - probably related directly to OC/ heat, while HW errors is NMW + .. comm errors.. etc?

Send work off, get an answer. The answer should be a valid share, and cgminer can track which chip bank it was sent to. If the response is not a valid share, it's an "invalid nonce" type of hardware error. If the response is garbled to the point of being gibberish and doesn't even look like a response, it's "no matching work" and cgminer can't even tell which chip was giving the garbage response. HW errors is the sum of both of these types of error.

Ah, ok thanks for the explanation. In short OC'n can cause both types.
So would each NMW constitute as an invalid diff1 share? cuz afaik HW errors are all equivilant to invalid diff1 shares .. which is why it used as 2% of diff1shares for --avalon-auto...?

In other words could one calculate the ratio of NMW to HW errors by doing (NMW / (HW errors - NMW)) * 100?

-ck

legendary

Activity: 4088

Merit: 1631

Ruu \o/

Quote from: GenTarkin on July 09, 2013, 03:56:16 PM

Quote from: -ck on July 09, 2013, 01:38:18 PM

Quote from: GenTarkin on July 09, 2013, 10:50:09 AM

Quote from: kano on July 09, 2013, 06:55:09 AM

I looked up the code (in relation to my HW error comments before)
HW is the count of bad nonces generated plus NMW
NMW is when work comes back but it doesn't match any work sent to it (thus cannot be used)

Anyway to minimize NMW?
My NMW hovers right around .52% of all matching work...

Or is NMW ALSO the result of chip errors?

Chip errors, even worse than regular HW errors.

Oh, so let me see if I understand, NMW is just chip errors - probably related directly to OC/ heat, while HW errors is NMW + .. comm errors.. etc?

Send work off, get an answer. The answer should be a valid share, and cgminer can track which chip bank it was sent to. If the response is not a valid share, it's an "invalid nonce" type of hardware error. If the response is garbled to the point of being gibberish and doesn't even look like a response, it's "no matching work" and cgminer can't even tell which chip was giving the garbage response. HW errors is the sum of both of these types of error.

GenTarkin

legendary

Activity: 2450

Merit: 1002

Quote from: -ck on July 09, 2013, 01:38:18 PM

Quote from: GenTarkin on July 09, 2013, 10:50:09 AM

Quote from: kano on July 09, 2013, 06:55:09 AM

I looked up the code (in relation to my HW error comments before)
HW is the count of bad nonces generated plus NMW
NMW is when work comes back but it doesn't match any work sent to it (thus cannot be used)

Anyway to minimize NMW?
My NMW hovers right around .52% of all matching work...

Or is NMW ALSO the result of chip errors?

Chip errors, even worse than regular HW errors.

Oh, so let me see if I understand, NMW is just chip errors - probably related directly to OC/ heat, while HW errors is NMW + .. comm errors.. etc?
OR
is HW errors a direct result of NMW?

Does NMW in any way relate to shares / share difficulty?

silverston

hero member

Activity: 952

Merit: 502

SAPG Pre-Sale Live on Uniswap!

dopamine
thanks man, all work fine now Wink

-ck

legendary

Activity: 4088

Merit: 1631

Ruu \o/

Quote from: sneef on July 09, 2013, 03:14:03 PM

Is there a user manual page for the --options that we can run, i only found --avalon-fan by stumbling on someone else's post.

I am running 0703 firmware from ckolivas (thank you Sir).

https://raw.github.com/ckolivas/cgminer/master/ASIC-README

mgio

hero member

Activity: 546

Merit: 500

Quote from: cypherdoc on July 09, 2013, 12:36:18 PM

Quote from: mgio on July 09, 2013, 02:01:57 AM

Finally got my Avalon today!

I'm running it with dynamic speed adjustment and I have it on a desk and a box right in front of my window AC unit which is feeding into the inlet. I have the AC set on high, cool, and 69 degrees. Seems right now frequency is stabilized at about 349MHz so far, and I'm getting about 87GH/s although sometimes it's as high as 90 or as low as 83. Probably needs more time to find out what the real average is.

I have a better AC unit coming tomorrow. It's a standalone unit and I'm going to have it piped right into the avalon intake. It's 12000 BTU so it might be a bit of an overkill. Hopefully I can run it on low. Will I be able to get above 90GH/s or so without getting a new power supply?

Anyways, my question is why does it show as a much lower (78GH/s or so) hash rate on the 50BTC.com website. Should I be using a different pool perhaps (I signed up for ozcoin but haven't tried it yet)? I believe 50BTC.com supports the stratum protocol and it says "true" under the stratum entries on the cgminer table, except for the last one "StratumURL".

Which is the hardware error field, btw? And what should that be kept below? Ugh, sorry for all the questions. I'm reading up as many threads as I can about the Avalon now, but it's hard to find all this information in one place (yes, I've read the wiki).

which AC unit did you order? sounds like a good idea.

Here is a screenshot of my CGMiner status. I'm not sure which column is the hardware errors. Does everything else look good?

The AC I got is a Sylvia SYL-12PE.
See: http://www.amazon.com/Sylvania-Portable-Conditioner-Dehumidifer-SYL-12PE/dp/B008J5CFOM/ref=cm_cr_pr_product_top

I think it draws 1000+ watts when the compressor is running, so I'm worried about tripping a breaker even if the compressor doesn't run very often. I would have bought a 8000 BTU unit but I bought this off a friend for $200 so I figure it was a good deal.

I'm going to home depot this evening and get some supplies to build some ductwork for it.

Topic: Avalon ASIC users thread - page 126. (Read 438516 times)