you da man....
I am stuck on windows with untill I can get linuxcoin to cooperate.
at 1125/1000 I11 672Mh I am happy with that.. I have not tried to clock it up.. think I will try that now.
Topic: collection for cgminer 7970 [Card as been sent! THANK YOU EVERYONE] (Read 9935 times)
March 31, 2012, 12:20:08 AM
cgminer 2.3.2 is out with a new poclbm kernel I've been bashing with a mallet for a week to try and extract some more out of it, and I hit my target which is 720 MHash at 1200/1050+5% clocks intensity 11 with 12.3 amd drivers. 
February 20, 2012, 12:27:03 AM
Cgminer 2.2.7 out. This version fixes the bug with 12.2 ATI drivers. Reworked to use -w 64 by default on Tahiti which is worth just under 1 more MHash. 718.5 MHash now at 1200/1050+5% clocks intensity 11.
February 16, 2012, 05:19:27 AM
cgminer 2.2.6 out. 2+ more mhash on 7970. It's getting harder and harder to extract much more 
1200/1050+5% clocks, intensity 11 - 717 Mhash.
I'm going to have to look at what methods you're using. I'm curious as to how the programming differs between VLIW and GCN.1200/1050+5% clocks, intensity 11 - 717 Mhash.
https://github.com/ckolivas/cgminer/blob/master/poclbm120214.cl
It's using a worksize of 256, 1 vector (i.e. no vectors) and intensity 11.
February 16, 2012, 05:17:42 AM
cgminer 2.2.6 out. 2+ more mhash on 7970. It's getting harder and harder to extract much more
1200/1050+5% clocks, intensity 11 - 717 Mhash.
I'm going to have to look at what methods you're using. I'm curious as to how the programming differs between VLIW and GCN. 1200/1050+5% clocks, intensity 11 - 717 Mhash.
February 15, 2012, 05:49:32 PM
cgminer 2.2.6 out. 2+ more mhash on 7970. It's getting harder and harder to extract much more
1200/1050+5% clocks, intensity 11 - 717 Mhash.
1200/1050+5% clocks, intensity 11 - 717 Mhash.
February 14, 2012, 08:54:41 AM
The problem, I believe, is that the GCN cards are built to take large vector counts and perform a single instruction on them all at once. This is in contrast to the small vector count of VLIW which can perform large instructions quickly. So, a straight-forward 16-vector miner with simple instructions should work better than a 4-vector miner with multiple instructions. Granted, this is speculation, but from what I've seen of the hardware specs, it should hold true.
I tried it, and a 16 vector simplest possible mining kernel had performance which was, unfortunately, appalling. GCN with SDK 2.6 (the only one it works with) really does not want any vectors at all.If you have any documentation, I would really appreciate it.
February 13, 2012, 05:28:47 AM
The problem, I believe, is that the GCN cards are built to take large vector counts and perform a single instruction on them all at once. This is in contrast to the small vector count of VLIW which can perform large instructions quickly. So, a straight-forward 16-vector miner with simple instructions should work better than a 4-vector miner with multiple instructions. Granted, this is speculation, but from what I've seen of the hardware specs, it should hold true.
I tried it, and a 16 vector simplest possible mining kernel had performance which was, unfortunately, appalling. GCN with SDK 2.6 (the only one it works with) really does not want any vectors at all. February 13, 2012, 05:00:52 AM
The problem, I believe, is that the GCN cards are built to take large vector counts and perform a single instruction on them all at once. This is in contrast to the small vector count of VLIW which can perform large instructions quickly. So, a straight-forward 16-vector miner with simple instructions should work better than a 4-vector miner with multiple instructions. Granted, this is speculation, but from what I've seen of the hardware specs, it should hold true.
I tried it, and a 16 vector simplest possible mining kernel had performance which was, unfortunately, appalling. GCN with SDK 2.6 (the only one it works with) really does not want any vectors at all. February 13, 2012, 01:17:59 AM
The problem, I believe, is that the GCN cards are built to take large vector counts and perform a single instruction on them all at once. This is in contrast to the small vector count of VLIW which can perform large instructions quickly. So, a straight-forward 16-vector miner with simple instructions should work better than a 4-vector miner with multiple instructions. Granted, this is speculation, but from what I've seen of the hardware specs, it should hold true.
I tried it, and a 16 vector simplest possible mining kernel had performance which was, unfortunately, appalling. GCN with SDK 2.6 (the only one it works with) really does not want any vectors at all. February 13, 2012, 01:16:45 AM
New release of cgminer 2.2.5 with fresh kernel and no more zero binary error, coping with multiple different cards at last and working well with sdk 2.6.
7970 running at 1200/1050+5% is getting 714 Mhash with -I 11 and cgminer 2.2.5 defaults.
7970 running at 1200/1050+5% is getting 714 Mhash with -I 11 and cgminer 2.2.5 defaults.
February 12, 2012, 07:16:26 AM
Working on the 7970 tuning, I have tried to port both the diapolo and diablo kernels to cgminer. Alas neither of them are actually working yet, so instead I started modifying the existing poclbm kernel in cgminer to improve throughput. This should work on other GPUs as well as the 7970, but I have no idea if it's better or worse than phatk.
When it's released it will get a new date/version number, but I haven't changed the number right now so that people can download it now and give it a try:
https://raw.github.com/ckolivas/cgminer/kernels/poclbm120203.cl
Remember to delete any .bin files and if you're not on a 7970 with the latest cgminer, you'll have to tell it to use that kernel with -k poclbm.
So what's the damage? Well on the 7970 at 1200/1050 clocks, which was getting 694MHash, it's now getting 711Mhash. The 7970 has this unusual behaviour where the hashrate slowly rises for the first 5-10 minutes.
When it's released it will get a new date/version number, but I haven't changed the number right now so that people can download it now and give it a try:
https://raw.github.com/ckolivas/cgminer/kernels/poclbm120203.cl
Remember to delete any .bin files and if you're not on a 7970 with the latest cgminer, you'll have to tell it to use that kernel with -k poclbm.
So what's the damage? Well on the 7970 at 1200/1050 clocks, which was getting 694MHash, it's now getting 711Mhash. The 7970 has this unusual behaviour where the hashrate slowly rises for the first 5-10 minutes.
February 10, 2012, 10:35:07 AM
Working on the 7970 tuning, I have tried to port both the diapolo and diablo kernels to cgminer. Alas neither of them are actually working yet, so instead I started modifying the existing poclbm kernel in cgminer to improve throughput. This should work on other GPUs as well as the 7970, but I have no idea if it's better or worse than phatk.
When it's released it will get a new date/version number, but I haven't changed the number right now so that people can download it now and give it a try:
https://raw.github.com/ckolivas/cgminer/kernels/poclbm120203.cl
Remember to delete any .bin files and if you're not on a 7970 with the latest cgminer, you'll have to tell it to use that kernel with -k poclbm.
So what's the damage? Well on the 7970 at 1200/1050 clocks, which was getting 694MHash, it's now getting 711Mhash. The 7970 has this unusual behaviour where the hashrate slowly rises for the first 5-10 minutes.
When it's released it will get a new date/version number, but I haven't changed the number right now so that people can download it now and give it a try:
https://raw.github.com/ckolivas/cgminer/kernels/poclbm120203.cl
Remember to delete any .bin files and if you're not on a 7970 with the latest cgminer, you'll have to tell it to use that kernel with -k poclbm.
So what's the damage? Well on the 7970 at 1200/1050 clocks, which was getting 694MHash, it's now getting 711Mhash. The 7970 has this unusual behaviour where the hashrate slowly rises for the first 5-10 minutes.
February 08, 2012, 10:33:39 PM
Many thanks to ck the hard work and everyone else who donated!
I just changed over to 2.2.3 on my windows machine with two shiny new Sapphire 7970s and am getting 513 and 537 at stock stock speeds (running at a leisurely 78 degrees), a huge jump over 2.2.1.
Many thanks!
Excellent. Don't forget to use intensity 11 if you can.I just changed over to 2.2.3 on my windows machine with two shiny new Sapphire 7970s and am getting 513 and 537 at stock stock speeds (running at a leisurely 78 degrees), a huge jump over 2.2.1.
Many thanks!
I use dynamic on my main card and have been using 14 on the second as it's a Crossfire only card (I haven't progressed to the whole eyefinity 2 x 6 thang yet
). I've changed to 11 and it seems to have basically the same rate. Coolness! February 08, 2012, 09:52:59 PM
Many thanks to ck the hard work and everyone else who donated!
I just changed over to 2.2.3 on my windows machine with two shiny new Sapphire 7970s and am getting 513 and 537 at stock stock speeds (running at a leisurely 78 degrees), a huge jump over 2.2.1.
Many thanks!
Excellent. Don't forget to use intensity 11 if you can. I just changed over to 2.2.3 on my windows machine with two shiny new Sapphire 7970s and am getting 513 and 537 at stock stock speeds (running at a leisurely 78 degrees), a huge jump over 2.2.1.
Many thanks!
February 08, 2012, 09:50:46 PM
Many thanks to ck the hard work and everyone else who donated!
I just changed over to 2.2.3 on my windows machine with two shiny new Sapphire 7970s and am getting 513 and 537 at stock stock speeds (running at a leisurely 78 degrees), a huge jump over 2.2.1.
Many thanks!
I just changed over to 2.2.3 on my windows machine with two shiny new Sapphire 7970s and am getting 513 and 537 at stock stock speeds (running at a leisurely 78 degrees), a huge jump over 2.2.1.
Many thanks!
February 08, 2012, 06:41:42 PM
Looks like I will have to figure out how to get this thing working on linuxcoin...
I imagine it is something like
curl somespecialfile
edit xorg.conf directly
aticonfig initialize
?
https://bitcointalksearch.org/topic/m.738468
That has the basics.
February 08, 2012, 06:30:09 PM
Looks like I will have to figure out how to get this thing working on linuxcoin...
I imagine it is something like
curl somespecialfile
edit xorg.conf directly
aticonfig initialize
?
February 08, 2012, 05:30:46 PM
And just for completeness, after the intensity investigation, I bumped it back up to the 1200/1050(+5%) clockspeed and put intensity up to 11 which seems the sweet spot. The final performance after it stabilised is:
Code:
GPU 0: 72.0C 4324RPM | 695.1/695.3Mh/s | A:98 R:1 HW:0 U: 9.79/m I:11
February 08, 2012, 05:18:04 PM
Ck: can you do the comparisons with the clock not set so high. at 1200 they run pretty hot with a high fan, also if you just do down about 50 on the clock, you can reduce the voltage down to 1120 or so.
I am currently running at volt: 1100 clock: 1100 650 Mhash diablo
At your clock settings of 1100 (+5% powertune) I'm getting 638 Mhash. Voltage doesn't change on linux with ADL it seems.I am currently running at volt: 1100 clock: 1100 650 Mhash diablo
Doing:
Code:
export GPU_USE_SYNC_OBJECTS=1
was mandatory for getting rid of the CPU usage bug. Interestingly with that feature enabled, there is an inflexion point in intensity when the CPU usage jumps up dramatically. With the cgminer default of 2 GPU threads, it is low CPU usage up to intensity 11. With 1 GPU thread, it is low CPU usage up to intensity 13. Both seem to provide only about 1MH difference.Jump to: