ROCKMINER ASIC miner official thread - page 10.

AJRGale

hero member

Activity: 767

Merit: 500

Quote from: -ck on November 18, 2014, 05:11:00 AM

Quote from: AJRGale on November 18, 2014, 05:02:52 AM

Quote from: -ck on November 18, 2014, 04:17:57 AM

Quote from: AJRGale on November 18, 2014, 04:01:35 AM

Quote from: AJRGale on November 18, 2014, 02:38:57 AM

right'o, recompiled a new git download..

i'll let it run for a few hours to balance out and i'll post results

well WTF did i do wrong?

No idea, your results look virtually unchanged (and still shitful). You definitely built the latest from git? What does 'git describe' say?

10GH/s less again at about 40% errors be it issue
"git describe" give me : v1.0-7211-g38700f5 if that means anything to you?

Okay, that's what we in the industry call "fucking shithouse". No idea what's wrong sorry.

This is how I see 8 of them perform on an extended run on ckpool; i.e. not religius (remote login, not my hardware):

Code:

A:76004520 R:1238160 HW:855006 WU:12408.0/m

0: LIR 0 : 350MHz 43C | 101.4G / 107.5Gh/s WU:1502.4/m
1: LIR 1 : 350MHz 42C | 112.7G / 112.8Gh/s WU:1575.8/m
2: LIR 2 : 350MHz 40C | 112.7G / 116.7Gh/s WU:1630.8/m
3: LIR 3 : 350MHz 42C | 123.4G / 113.5Gh/s WU:1585.1/m
4: LIR 4 : 350MHz 40C | 119.2G / 109.3Gh/s WU:1526.5/m
5: LIR 5 : 350MHz 42C | 106.3G / 107.2Gh/s WU:1497.0/m
6: LIR 6 : 350MHz 41C | 108.0G / 106.5Gh/s WU:1487.5/m
7: LIR 7 : 350MHz 42C | 114.2G / 114.7Gh/s WU:1603.0/m

Perhaps you have a hardware issue?

im asuming so, when i get the parts, im modding the heatsinks in the thing, the top heatsink when i popped it off, 3 or 4 of the chips were not making contact with the gunk and sink., the bottom side has more gunk then i've ever used in the industry of computer repairs (who knows how many air bubbles are in it). i'd like to see what each chip is spitting out with its temp and hashing. but your software doesn't show that out of the box.
(lukes does, and atm does give better hash-rates (80-85Ghs) but i cant be stuffed right now installing it on the 'buntu system)

all said an done, i'll just accept its a shit model i got. the best i got out of it was 100Ghs, with 50% errors (how is the errors counted anyway? is it just garbage that the software picks up, or is it the hardware going "ouchies, i subbed my toe") on your software.

Ntrain2k

hero member

Activity: 527

Merit: 500

HiveNet - Distributed Cloud Computing

Unfortunately the win exe didn't work for me either.

I was topping out around 400 gh with 7 new r-boxes and 8 of the old ones.

-ck

legendary

Activity: 4088

Merit: 1631

Ruu \o/

Quote from: AJRGale on November 18, 2014, 05:02:52 AM

Quote from: -ck on November 18, 2014, 04:17:57 AM

Quote from: AJRGale on November 18, 2014, 04:01:35 AM

Quote from: AJRGale on November 18, 2014, 02:38:57 AM

right'o, recompiled a new git download..

i'll let it run for a few hours to balance out and i'll post results

well WTF did i do wrong?

No idea, your results look virtually unchanged (and still shitful). You definitely built the latest from git? What does 'git describe' say?

10GH/s less again at about 40% errors be it issue
"git describe" give me : v1.0-7211-g38700f5 if that means anything to you?

Okay, that's what we in the industry call "fucking shithouse". No idea what's wrong sorry.

This is how I see 8 of them perform on an extended run on ckpool; i.e. not religius (remote login, not my hardware):

Code:

A:76004520 R:1238160 HW:855006 WU:12408.0/m

0: LIR 0 : 350MHz 43C | 101.4G / 107.5Gh/s WU:1502.4/m
1: LIR 1 : 350MHz 42C | 112.7G / 112.8Gh/s WU:1575.8/m
2: LIR 2 : 350MHz 40C | 112.7G / 116.7Gh/s WU:1630.8/m
3: LIR 3 : 350MHz 42C | 123.4G / 113.5Gh/s WU:1585.1/m
4: LIR 4 : 350MHz 40C | 119.2G / 109.3Gh/s WU:1526.5/m
5: LIR 5 : 350MHz 42C | 106.3G / 107.2Gh/s WU:1497.0/m
6: LIR 6 : 350MHz 41C | 108.0G / 106.5Gh/s WU:1487.5/m
7: LIR 7 : 350MHz 42C | 114.2G / 114.7Gh/s WU:1603.0/m

Perhaps you have a hardware issue?

AJRGale

hero member

Activity: 767

Merit: 500

Quote from: -ck on November 18, 2014, 04:17:57 AM

Quote from: AJRGale on November 18, 2014, 04:01:35 AM

Quote from: AJRGale on November 18, 2014, 02:38:57 AM

right'o, recompiled a new git download..

i'll let it run for a few hours to balance out and i'll post results

well WTF did i do wrong?

No idea, your results look virtually unchanged (and still shitful). You definitely built the latest from git? What does 'git describe' say?

10GH/s less again at about 40% errors be it issue
"git describe" give me : v1.0-7211-g38700f5 if that means anything to you?

-ck

legendary

Activity: 4088

Merit: 1631

Ruu \o/

Quote from: AJRGale on November 18, 2014, 04:01:35 AM

Quote from: AJRGale on November 18, 2014, 02:38:57 AM

right'o, recompiled a new git download..

i'll let it run for a few hours to balance out and i'll post results

well WTF did i do wrong?

No idea, your results look virtually unchanged (and still shitful). You definitely built the latest from git? What does 'git describe' say?

AJRGale

hero member

Activity: 767

Merit: 500

Quote from: AJRGale on November 18, 2014, 02:38:57 AM

right'o, recompiled a new git download..

i'll let it run for a few hours to balance out and i'll post results

well WTF did i do wrong?

AJRGale

hero member

Activity: 767

Merit: 500

right'o, recompiled a new git download..

i'll let it run for a few hours to balance out and i'll post results

-ck

legendary

Activity: 4088

Merit: 1631

Ruu \o/

Quote from: AJRGale on November 18, 2014, 02:11:14 AM

Well, running on 'buntu, with 4.7.1: here is the result:

running slower (70-80GH/s over 100GH/s), and ~10% hardware errors!

now, if i was to run it at 330MHz, its 60GH/s

Try latest git instead. The rockminer driver hadn't received any attention recently till after 4.7.1

AJRGale

hero member

Activity: 767

Merit: 500

Well, running on 'buntu, with 4.7.1: here is the result:

running slower (70-80GH/s over 100GH/s), and ~10% hardware errors!

now, if i was to run it at 330MHz, its 60GH/s

AllInBit

member

Activity: 60

Merit: 10

Quote from: elasticband on November 12, 2014, 01:48:50 PM

Quote from: AllInBit on November 12, 2014, 01:40:55 PM

Quote from: rockxie on October 20, 2014, 11:14:21 PM

Quote from: AllInBit on October 20, 2014, 07:23:44 PM

Is there a Sep. financial report/update coming soon? Thanks

We will combine September and October report together by considerring Oct has passed two third.

Any updates??

basically no updates or information for september or october.

Maybe an update on RM's involvment with AMhash? what revenue is the company pulling in from this endeavour? Is RM personally self mining? whats the hashrate or why not?
Are RM just taking a back seat until AMgen4?

Quote

September 13th, 2014 - Update and Dividends of August can be visited at : http://blog.rockminer.com/t/7/

July 28th, 2014 - Financial report of ROCKMINER will be opened on 29th July.

ROCKMINER TEAM

June 26th, 2014 - the ROCKMINER (ROCK) Fund Will be open for trading on HavelockInvestments.com starting Friday at 9:00 A.M GMT

Thank you,

Rockminer
Havelock Investments

Looks like FC gave Rockxie the low down on how to handle PR. Just ignore them!

Ntrain2k

hero member

Activity: 527

Merit: 500

HiveNet - Distributed Cloud Computing

Quote from: -ck on November 13, 2014, 09:05:51 PM

Quote from: AJRGale on November 13, 2014, 07:31:10 PM

hate to be an ass, but what do you mean buy "it's at the diminishing returns stage"? the more you tweak the software around the less its giving you back? or less its worth it to you? or less its worth the r/rockets-boxes to be running?

Each next tweak is improving things less and less. Barring some kind of breakthrough (like getting some goddamned documentation about the communication protocol from the manufacturer instead of using their half working driver as an example) I'm not sure things stand to improve much more.

Quote from: Ntrain2k on November 13, 2014, 07:39:46 PM

If there an updated win version yet?

Heh can't seem to keep with the demands on this place... but yeah sure here's a binary:

http://ck.kolivas.org/apps/cgminer/temp/cgminer.exe

Just drop it into an extracted 4.7.1 directory overwriting the existing binary.

Thank you Sir. I'll give it a shot.

AJRGale

hero member

Activity: 767

Merit: 500

.. actually, is Rockxie going to pop back in here again? last message was on October 21, about 3 weeks ago

all i can see is he'd be all up in that amhash thing (or the other asicminer reinvent meh thingy)

-ck

legendary

Activity: 4088

Merit: 1631

Ruu \o/

Quote from: AJRGale on November 13, 2014, 07:31:10 PM

hate to be an ass, but what do you mean buy "it's at the diminishing returns stage"? the more you tweak the software around the less its giving you back? or less its worth it to you? or less its worth the r/rockets-boxes to be running?

Each next tweak is improving things less and less. Barring some kind of breakthrough (like getting some goddamned documentation about the communication protocol from the manufacturer instead of using their half working driver as an example) I'm not sure things stand to improve much more.

Quote from: Ntrain2k on November 13, 2014, 07:39:46 PM

If there an updated win version yet?

Heh can't seem to keep with the demands on this place... but yeah sure here's a binary:

http://ck.kolivas.org/apps/cgminer/temp/cgminer.exe

Just drop it into an extracted 4.7.1 directory overwriting the existing binary.

AJRGale

hero member

Activity: 767

Merit: 500

Quote from: MilkyLep on November 13, 2014, 07:36:05 PM

Quote from: AJRGale on November 13, 2014, 07:31:10 PM

Quote from: -ck on November 13, 2014, 07:01:41 PM

snipples

Should work with both. The Rocketbox was 8 blades, the new Rbox is a single blade as Dogie said. I'm still trying to further tweak the driver, but it's at the diminishing returns stage. A blade seems to drop out very easily though, and it usually re-hotplugs but not always.

hate to be an ass, but what do you mean buy "it's at the diminishing returns stage"? the more you tweak the software around the less its giving you back? or less its worth it to you? or less its worth the r/rockets-boxes to be running?

Sounds like he's reaching a point where he won't be able to squeeze out additional efficiency/performance.

i suspect that. im just hoping the changes made for the rkbox is going to be enough for the nrbox, since mine is doing 50% Hardware Errors.
and no, the "official" cgminer off the suppliers site does not work, it errors out, 4.7.0 works, with 50% errors and using Icarus drivers, (doesn't work without it.) and yes, ive tried a different cable, no change. the 4.7.1 i have compiled last night on a ubuntu laptop should change that, but i wouldn't be surprised it wont. even BFGminer is displaying the same error rate.

as i mentioned before i suspect the heatsink is to blame, there is little warmth coming out of it, even though the system says its 15/20 degrees over the ambient, i should feel that coming out of it.

Ntrain2k

hero member

Activity: 527

Merit: 500

HiveNet - Distributed Cloud Computing

Quote from: -ck on November 13, 2014, 07:14:10 PM

Quote from: Ntrain2k on November 13, 2014, 07:04:15 PM

Any idea why I can't get my r-boxes to run at full speed with cg?

Cg does not see my miners on my other box either.

Quote from: -ck on November 13, 2014, 05:10:05 AM

Note the displayed hashrate appears lower but that's because the dodgy driver counts rejects in its hashrate which there are quite a few in the form of duplicates, but the updated driver hashrate is higher as confirmed at the pool end.

No idea about your other box. USB3 slot maybe?

If there an updated win version yet?

I'm using USB2 on the box that won't pick them up on. System sees them fine with the regular driver and BFG.

MilkyLep

hero member

Activity: 583

Merit: 500

Bitcoin for all & all for Bitcoin

Quote from: AJRGale on November 13, 2014, 07:31:10 PM

Quote from: -ck on November 13, 2014, 07:01:41 PM

snipples

Should work with both. The Rocketbox was 8 blades, the new Rbox is a single blade as Dogie said. I'm still trying to further tweak the driver, but it's at the diminishing returns stage. A blade seems to drop out very easily though, and it usually re-hotplugs but not always.

hate to be an ass, but what do you mean buy "it's at the diminishing returns stage"? the more you tweak the software around the less its giving you back? or less its worth it to you? or less its worth the r/rockets-boxes to be running?

Sounds like he's reaching a point where he won't be able to squeeze out additional efficiency/performance.

AJRGale

hero member

Activity: 767

Merit: 500

Quote from: -ck on November 13, 2014, 07:01:41 PM

snipples

Should work with both. The Rocketbox was 8 blades, the new Rbox is a single blade as Dogie said. I'm still trying to further tweak the driver, but it's at the diminishing returns stage. A blade seems to drop out very easily though, and it usually re-hotplugs but not always.

hate to be an ass, but what do you mean buy "it's at the diminishing returns stage"? the more you tweak the software around the less its giving you back? or less its worth it to you? or less its worth the r/rockets-boxes to be running?

Trends

member

Activity: 119

Merit: 10

Quote from: Trends on November 03, 2014, 05:49:50 PM

ROCKMINERS Research & Development Theory

" LET THE MINERS FIGURE IT OUT"

AJRGale

hero member

Activity: 767

Merit: 500

Quote from: dogie on November 13, 2014, 06:55:16 PM

Quote from: AJRGale on November 13, 2014, 06:24:21 PM

Quote from: hurricandave on November 13, 2014, 06:16:29 PM

Quote from: AJRGale on November 13, 2014, 05:57:19 PM

Hmm, RKBox equivalent to the "New Rbox"? also just git'ed cgminer last night into a ubuntu system, setting that up for the box, gonna see if that works.

Depends on what you mean by equivalent? But yes these two particular devices use the same version of software/drivers.

Equivalent as in the same chipset, firmware, etc

But if the mucking around with the RKbox that ckolivas did was enough for the software to work nicely with the "New Rbox" that guy deserves a beer or 20

New Rbox (pics) is a single blade from an RK / R3 box (pics).

So they are different: The RK the controller is a separate board, connected by 2x 16 pin jumpers, don't know what the chip is, or the firmware is. and its using the BE200 Revision: 9P01A. the silkscreen has "ROCMINER_BLADE_V1.2 2014_06_02 The "New Rbox" is using an Atmel mega88pa, soldered onto the board, and using the BE200 revision: 9P02A, and its Silk screen has: ROCKMINER_R-BOX_V2 2014_08_9 (odd, admitted the 0 before the 9? confirmed halflife2 ep 3?) they are almost the same. but different enough to have an affect.

-ck

legendary

Activity: 4088

Merit: 1631

Ruu \o/

Quote from: Ntrain2k on November 13, 2014, 07:04:15 PM

Any idea why I can't get my r-boxes to run at full speed with cg?

Cg does not see my miners on my other box either.

Quote from: -ck on November 13, 2014, 05:10:05 AM

Note the displayed hashrate appears lower but that's because the dodgy driver counts rejects in its hashrate which there are quite a few in the form of duplicates, but the updated driver hashrate is higher as confirmed at the pool end.

Also you didn't specify if you're using 4.7.1 or the latest code. Prior to yesterday, I had never looked at the rocketbox / new rbox code before.

No idea about your other box. USB3 slot maybe?

Topic: ROCKMINER ASIC miner official thread - page 10. (Read 199664 times)