Unofficial "High MHash"/"Best Settings" GPU Comparision Thread

[Mabsark]

Those are fake reports of guiminer I guess. Best to check your real hashrate is to use 24h stats on ozcoin. And could be as well that you're getting extra mhashes out of your cpu. Check your cpu usage.

Been there, done that. The results are legit. I've been using those settings for over 6 months and both guiminer and every pool I've mined at show the same results. I can't really check results over a 24 hour period as I'm not a dedicated miner, I'm a dedicated gamer.

Run a second worker on phoenix 1, and you get a major boost in Mh/s. Well, you do with 6870's anyway and someone else told me that they got a 40 Mh/s increase on a 5830 at stock settings.

[1714843261]

1714843261

[BinaryMage]

Those are fake reports of guiminer I guess. Best to check your real hashrate is to use 24h stats on ozcoin. And could be as well that you're getting extra mhashes out of your cpu. Check your cpu usage.

Been there, done that. The results are legit. I've been using those settings for over 6 months and both guiminer and every pool I've mined at show the same results. I can't really check results over a 24 hour period as I'm not a dedicated miner, I'm a dedicated gamer.

Run a second worker on phoenix 1, and you get a major boost in Mh/s. Well, you do with 6870's anyway and someone else told me that they got a 40 Mh/s increase on a 5830 at stock settings.

Just for verification, would you mind posting a screenshot? Not that I don't trust you, but those results are beyond anything I've ever seen for a 6870.

[-ck]

6870 Northern Islands has 1120 stream processors of the same architecture as the 5870 which has 1600. Performance is linearly related to architecture * stream processors * Mhz. Therefore at exact same clocks, the 6870 should be (1120/1600) 0.7 * Mhash of a 5870. At 1000Mhz, the 5870 gets at best ~460Mhash. Therefore a number as high as 375 for a 6870 is ... highly unlikely.

[Mabsark]

Just for verification, would you mind posting a screenshot? Not that I don't trust you, but those results are beyond anything I've ever seen for a 6870.

Here's some I posted in another thread the other day:





And here's another one I just took:

[Mabsark]

6870 Northern Islands has 1120 stream processors of the same architecture as the 5870 which has 1600. Performance is linearly related to architecture * stream processors * Mhz. Therefore at exact same clocks, the 6870 should be (1120/1600) 0.7 * Mhash of a 5870. At 1000Mhz, the 5870 gets at best ~460Mhash. Therefore a number as high as 375 for a 6870 is ... highly unlikely.

What makes you think that a 5870 can only get ~460 MH/s? Have you tried running 2 workers on a single 5870? Maybe you'll get over 500 MH/s if you do.

[-ck]

6870 Northern Islands has 1120 stream processors of the same architecture as the 5870 which has 1600. Performance is linearly related to architecture * stream processors * Mhz. Therefore at exact same clocks, the 6870 should be (1120/1600) 0.7 * Mhash of a 5870. At 1000Mhz, the 5870 gets at best ~460Mhash. Therefore a number as high as 375 for a 6870 is ... highly unlikely.

What makes you think that a 5870 can only get ~460 MH/s? Have you tried running 2 workers on a single 5870? Maybe you'll get over 500 MH/s if you do.

I'm sorry, clearly I don't know what I'm talking about.

[Mabsark]

I'm sorry, clearly I don't know what I'm talking about.

I'm not saying you don't know what you're talking about, I'm asking if you've tried running 2 workers on a single 5870 using Phoenix and phatk2.

[kano]

Just for verification, would you mind posting a screenshot? Not that I don't trust you, but those results are beyond anything I've ever seen for a 6870.

Here's some I posted in another thread the other day:

https://i.imgur.com/jr5jy.png

https://i.imgur.com/x6T1I.png

And here's another one I just took:

Interesting so when my Icarus ran for 3.5 hours at a share submission rate of 5.60 shares a minute I can take that as proof that an Icarus can hash at 401MH/s ?

No my Icarus (with default bitstream) can hash at a MAX of ~380MH/s

At 5.4 shares/minute 400 shares is only 74minutes ....

You seriously need to run for 24hours or your figures are quite MEANINGLESS.
Even at 24h it's possible to get a variance high enough to make it look better than it is.

[Mabsark]

You seriously need to run for 24hours or your figures are quite MEANINGLESS.
Even at 24h it's possible to get a variance high enough to make it look better than it is.

I can't, I've got too many games to play and not enough time to play them. Sure, variance can make it look higher (or worse) but by around 50 MH/s on a consistent basis ever since Catayst 11.7 came out? I find that highly unlikely.

[kano]

You seriously need to run for 24hours or your figures are quite MEANINGLESS.
Even at 24h it's possible to get a variance high enough to make it look better than it is.

I can't, I've got too many games to play and not enough time to play them. Sure, variance can make it look higher (or worse) but by around 50 MH/s on a consistent basis ever since Catayst 11.7 came out? I find that highly unlikely.

Shares/minute (submitted - not calculated)

[Mabsark]

Shares/minute (submitted - not calculated)

I not sure what trying to say. If you're asking for my shares/minute, then the answer is in the screenshots.

77 / 15 = 5.1333...
81 / 15 = 5.4

[kano]

Shares/minute (submitted - not calculated)

I not sure what trying to say. If you're asking for my shares/minute, then the answer is in the screenshots.

77 / 15 = 5.1333...
81 / 15 = 5.4

Yep as even that those 2 inconclusive short rounds show.

You keep stating that the MH/s was blah regularly - which of course you'd expect it to always be consistent.
However to determine if it might be accurate is to ether look at a long term shares/minute or look at the shares/minute you get regularly over time.

As that one shows there, it varies by 5% easily (but it can vary a lot more over short a period of time)

[Mabsark]

You keep stating that the MH/s was blah regularly - which of course you'd expect it to always be consistent.
However to determine if it might be accurate is to ether look at a long term shares/minute or look at the shares/minute you get regularly over time.

As that one shows there, it varies by 5% easily (but it can vary a lot more over short a period of time)

Like I said, I've been running with these settings since Catalyst 11.7 came out. It was that update which gave the increase. I already had 2 workers mining seperate pools at a total hashrate of ~315 MH/s and after that update my hashrate shot up. I've already determined that it's accurate, I don't need to do so again. That is my hashrate and it is due to the second worker.

If you don't want to believe that, then that's your choice. If anyone wants a possible extra 50 MH/s though, all they have to do is test this themselves. You don't have to take my word for it.

[kano]

I don't care what the hash rate says.
I'm not gonna take your word for it coz all you are doing is quoting what some program is telling you.

I care what the shares/minute say over time.
That can't lie. It has variance. As you just shown yourself with those 2 numbers.
But over a long period of time it will tell you what your hash rate really is.

[oldDirty]

If you don't want to believe that, then that's your choice. If anyone wants a possible extra 50 MH/s though, all they have to do is test this themselves. You don't have to take my word for it.

i try it and notice 10mh/s more for two 5850 850mhz. I can rise the gpu speed, but first i want to test this improvement for 24h.
Guiminer 2012-02-19

[Serenata]

Card Manufacturer: AMD
Card Model Number: HD 5850
Card OEM: SAPPHIRE (Xtreme)
MHash per Second: 301 (48 hours average)
Core Clock: 800
Memory Clock: 300
Average Operating Temperature: 74C
Ambient Temperature: 26C
Fan Speed: 47%
Host OS: Windows 7 64bit
Driver Version: Catalyst 12.4
Mining Program: cgminer 2.4.1
Command Line Flags/GUIminer Settings: diablo kernel, vectors 2, worksize 128, intensity 4, voltage 1.050

[conspirosphere.tk]

Here is my summer settings:
Card Model Number: HD 5870
Card OEM: Powercolor
MHash per Second: 378
Core Clock: 830
Memory Clock: 300
Average Operating Temperature: 64C
Ambient Temperature: Around 25C
Fan Speed: 40%
Host OS: Windows Seven
Driver Version: Catalyst 11.12 
Mining Program: Phoenix + AOCLBF
Command Line Flags/GUIminer Settings: phatk2 kernel, vectors, worksize 256, intensity 10, voltage 0.960

Card Model Number: HD 5830
Card OEM: XFX
MHash per Second: 230
Core Clock: 730
Memory Clock: 200
Average Operating Temperature: 63C
Ambient Temperature: Around 25C
Fan Speed: 46%
Host OS: Seven
Driver Version: Catalyst 11.12 
Mining Program: Phoenix + AOCLBF
Command Line Flags/GUIminer Settings: phatk2 kernel, vectors, worksize 256, intensity 10, voltage 0.950

[oldDirty]

Card Manufacturer: AMD x2
Card Model Number: HD 5850
Card OEM: MSI x2 OC Edition
MHash per Second: 330 (48 hours average)
Core Clock: 850
Memory Clock: 560
Average Operating Temperature: 64C
Ambient Temperature: 20C
Fan Speed: 37%
Host OS: Windows 7 64bit
Driver Version: Catalyst 12.3
Mining Program: cgminer 2.4.1
Command Line Flags/GUIminer Settings: diablo kernel, intensity gpu0=8, gpu1=14, voltage 1.024V

[BinaryMage]

Updated, thanks guys.

[stoppots]

If you don't want to believe that, then that's your choice. If anyone wants a possible extra 50 MH/s though, all they have to do is test this themselves. You don't have to take my word for it.

Mabsark, just so you know. You should really take another look at the post several people have made on the previous page and take their words for it. They have stated your conclusions are incorrect. They have also told you why they are incorrect. In addition to that you have told us in your own words how your reported numbers are completely incorrect and this is why the assumptions you have made with this data and believe to be true are in fact false.

With that said, if you are at all interested in knowing the true speed and potential earnings your hardware are capable of. You should really attempt to absorb the  comments and help those people have offered. If you go to either of the post made by ckolivas, and follow the link in his sig it will take you to the most widely used mining software miners are mining with. You may find it a little intimidating and a bit more in depth than guiminer, but it will not mislead you with false stats and will give you a more realistic view of wats really goin on when you are not playing games.

good luck and may the force be with you