AOA Forums AOA Forums AOA Forums Folding For Team 45 AOA Files Home Front Page Become an AOA Subscriber! UserCP Calendar Memberlist FAQ Search Forum Home


Go Back   AOA Forums > Hardware > AMD Motherboards & CPUs

AMD Motherboards & CPUs Questions or comments on AMD products?


Reply
 
LinkBack Thread Tools Rate Thread
  #1 (permalink)  
Old 29th June, 2006, 10:59 PM
gedon's Avatar
Member
 
Join Date: November 2004
Posts: 446

Is it a CPU-limit or...

It's like a joke. Several weeks ago I upgraded from an Asus A8N VM CSM to a DFI NF4 Ultra D.
My Opti 165 worked great at 1.4V and 2.65GHz on the Asus one. At 1.55V it didn't clock any MHz higher. I thought it was a board or chipset limitation.
Now, with an Ultra D I got 2.65GHz (295MHzx9) at 1.376V ("nearly stock volts"). Nice, but even with 1.55V 2.7GHz are not stable
Memory works great at 295MHz, after one week of priming, folding and encoding I can say, the system is 100% stable at 2.65GHz.

I tried nearly everything:
memorydivider (100, 133, 166), I don't know why, but the 180 divider is not stable
different LDT-, chipset- CPU-voltages
LDT multis 2.5, 3 and 4

I'll check CPU-multi/htt combos

bios-settings for 295MHzx9:
LDT x4
vLDT 1.30V
vChipset 1.50V
vcore 1.4V
mem 1:1
temp full load 40 degrees celsius

so, what do you think? Have you ever heard about such a limit?
I am confused.....
__________________
Digg this Post!Add Post to del.icio.usBookmark Post in TechnoratiFurl this Post!
Reply With Quote
  #2 (permalink)  
Old 29th June, 2006, 11:11 PM
Gizmo's Avatar
Chief BBS Administrator
BassTeroids Champion, Global Player Champion, Aim & Fire Champion, Puzzle Maniax Champion, Othello Champion, Canyon Glider Champion, Unicycle Challenge Champion, YetiSports 9: Final Spit Champion, Zed Champion
 
Join Date: May 2003
Location: Webb City, Mo
Posts: 16,178
Send a message via ICQ to Gizmo Send a message via AIM to Gizmo Send a message via MSN to Gizmo Send a message via Yahoo to Gizmo Send a message via Skype™ to Gizmo

All hardware has a limit. Pushing beyond that limit achieves little more than to make the hardware break.

Up until the 90 nM process came along, it was generally the case that putting in more volts got you more performance (at the expense of increased heat) in a more-or-less linear and direct relationship, right up to the point where things just stopped working altogether. As long as you could remove the heat, you were golden to keep pushing.

When we got to the 90 nM process though, things changed. We started seeing this a little bit at 130 nM, but at 90 nM it became VERY prevalent; leakage current. Leakage current is directly related to how much voltage there is in the circuit, and power is related to the square of current, so increasing voltage a little increases leakage significantly which increases power (and hence heat) a lot. As a result, with anything less than pretty extreme cooling, increasing voltage doesn't gain you much.

What HAS become evident, though (or at least, appears to be evident with the Intel process) is that at the 65 nM node, although voltage doesn't buy you much, keeping a chip cool can buy you a LOT.
Digg this Post!Add Post to del.icio.usBookmark Post in TechnoratiFurl this Post!
Reply With Quote
  #3 (permalink)  
Old 30th June, 2006, 01:34 AM
gedon's Avatar
Member
 
Join Date: November 2004
Posts: 446

thanx gizmo for the quick reply.
...leakage current... makes sense.
I hate those limits, you can do nothings to drop the temp down to 30 or less degrees with a watercooling. i would need a 100W Peltier element
__________________
Digg this Post!Add Post to del.icio.usBookmark Post in TechnoratiFurl this Post!
Reply With Quote
  #4 (permalink)  
Old 1st July, 2006, 07:03 PM
cloasters's Avatar
Asst. BBS Administrator
 
Join Date: September 2001
Location: Location, Location
Posts: 21,956

I guess using a Peltier would be easier and less expensive than refrigerating the cooling water? How much more speed would chilling the water down to say 33F or ~1C buy you?
__________________
When the world will be better.
Digg this Post!Add Post to del.icio.usBookmark Post in TechnoratiFurl this Post!
Reply With Quote
  #5 (permalink)  
Old 1st July, 2006, 07:20 PM
Member/Contributor/Resident Crystal Ball
 
Join Date: March 2004
Posts: 7,451

Sounds liek you've maxed the board, however temps may be part of the issue.

Try raising Max Async Latency and Read Preamble a bit to see if that gets you any higher. You might want to try re-mounting the cpu with a new thermal paste application as well...

If you still ahve the IHS attached, there made not be good contact under it. check the thermal offset using CBiD(Central Brain Identifier).

LDT should be 3x for 300htt!!!
__________________
Digg this Post!Add Post to del.icio.usBookmark Post in TechnoratiFurl this Post!
Reply With Quote
  #6 (permalink)  
Old 1st July, 2006, 07:25 PM
Gizmo's Avatar
Chief BBS Administrator
BassTeroids Champion, Global Player Champion, Aim & Fire Champion, Puzzle Maniax Champion, Othello Champion, Canyon Glider Champion, Unicycle Challenge Champion, YetiSports 9: Final Spit Champion, Zed Champion
 
Join Date: May 2003
Location: Webb City, Mo
Posts: 16,178
Send a message via ICQ to Gizmo Send a message via AIM to Gizmo Send a message via MSN to Gizmo Send a message via Yahoo to Gizmo Send a message via Skype™ to Gizmo

Quote:
Originally Posted by cloasters
I guess using a Peltier would be easier and less expensive than refrigerating the cooling water? How much more speed would chilling the water down to say 33F or ~1C buy you?
I don't know what the behavior is now, with these smaller processes, but at 130 nM, it was the case that performance scaled pretty linearly with absolute temperature, so that cutting the absolute temp in half would roughly double performance. Going from 40C to 1C would represent a change in absolute temp of about 13%, and thus should buy you about that much performance improvement, all other things being equal. However, the extra temp margin would then allow you to pump more power through the chip, providing additional noise margin in the signals, and allowing to push the chip harder.

With these new 65 nM chips, I dunno, but it sure seems like the linear temp/speed relationship has change dramatically.
Digg this Post!Add Post to del.icio.usBookmark Post in TechnoratiFurl this Post!
Reply With Quote
  #7 (permalink)  
Old 1st July, 2006, 07:50 PM
Member/Contributor/Resident Crystal Ball
 
Join Date: March 2004
Posts: 7,451

AM2's are not 65nm.....yet. 939's definately are not(90nm)
__________________
Digg this Post!Add Post to del.icio.usBookmark Post in TechnoratiFurl this Post!
Reply With Quote
  #8 (permalink)  
Old 1st July, 2006, 08:44 PM
Gizmo's Avatar
Chief BBS Administrator
BassTeroids Champion, Global Player Champion, Aim & Fire Champion, Puzzle Maniax Champion, Othello Champion, Canyon Glider Champion, Unicycle Challenge Champion, YetiSports 9: Final Spit Champion, Zed Champion
 
Join Date: May 2003
Location: Webb City, Mo
Posts: 16,178
Send a message via ICQ to Gizmo Send a message via AIM to Gizmo Send a message via MSN to Gizmo Send a message via Yahoo to Gizmo Send a message via Skype™ to Gizmo

True. I was thinking in terms of Conroe.
Digg this Post!Add Post to del.icio.usBookmark Post in TechnoratiFurl this Post!
Reply With Quote
  #9 (permalink)  
Old 1st July, 2006, 11:24 PM
Member/Contributor/Resident Crystal Ball
 
Join Date: March 2004
Posts: 7,451

A little pre-occupied are we?
Don't blame ya one bit.
__________________
Digg this Post!Add Post to del.icio.usBookmark Post in TechnoratiFurl this Post!
Reply With Quote
  #10 (permalink)  
Old 1st July, 2006, 11:46 PM
Gizmo's Avatar
Chief BBS Administrator
BassTeroids Champion, Global Player Champion, Aim & Fire Champion, Puzzle Maniax Champion, Othello Champion, Canyon Glider Champion, Unicycle Challenge Champion, YetiSports 9: Final Spit Champion, Zed Champion
 
Join Date: May 2003
Location: Webb City, Mo
Posts: 16,178
Send a message via ICQ to Gizmo Send a message via AIM to Gizmo Send a message via MSN to Gizmo Send a message via Yahoo to Gizmo Send a message via Skype™ to Gizmo

Quote:
Originally Posted by cadaveca
A little pre-occupied are we?
Don't blame ya one bit.
Err.....yeah.
Digg this Post!Add Post to del.icio.usBookmark Post in TechnoratiFurl this Post!
Reply With Quote
  #11 (permalink)  
Old 2nd July, 2006, 12:51 PM
gedon's Avatar
Member
 
Join Date: November 2004
Posts: 446

Quote:
Originally Posted by cadaveca
Sounds liek you've maxed the board, however temps may be part of the issue.

Try raising Max Async Latency and Read Preamble a bit to see if that gets you any higher. You might want to try re-mounting the cpu with a new thermal paste application as well...

If you still ahve the IHS attached, there made not be good contact under it. check the thermal offset using CBiD(Central Brain Identifier).

LDT should be 3x for 300htt!!!

I cut off the IHS yesterday morning. got 3-5K less, but no increase in clockspeed.
set max async latency and read preamble from 5/8 to 6/10. this works well. priming at 2808MHz, 1.504V, 40 degrees since 30min
can you explain why max async latency and read preamble influence the cpuclock?
__________________
Digg this Post!Add Post to del.icio.usBookmark Post in TechnoratiFurl this Post!
Reply With Quote
  #12 (permalink)  
Old 2nd July, 2006, 06:30 PM
Member/Contributor/Resident Crystal Ball
 
Join Date: March 2004
Posts: 7,451

Memory controller is on cpu die...directly interfaces with memory, so memory settings can greatly affect clocks, as they are also settings for cpu.
__________________
Digg this Post!Add Post to del.icio.usBookmark Post in TechnoratiFurl this Post!
Reply With Quote
  #13 (permalink)  
Old 2nd July, 2006, 07:18 PM
gedon's Avatar
Member
 
Join Date: November 2004
Posts: 446

now I'm up to 2808MHz (312MHzx9) and got the mem up to 280MHz. also with a 100 or 133 divider I can't push cpu any higher. OK, with 1.520V at 2.8GHz it's nearly at it's end. the rest is finetuning. I think 1.6V is possible. but at 2850MHz and 1.6V system crashes as it did before at 2695MHz, I mean bluescreens, restarts, system freezes. there is no rounding error in prime before system crashes.
what can I set for Max Async Latency and Read Preamble?
now at 10 and 6, maximum I can set is 15 and 9.5. Is it wise to continue raising up these two settings?
__________________
Digg this Post!Add Post to del.icio.usBookmark Post in TechnoratiFurl this Post!
Reply With Quote
  #14 (permalink)  
Old 2nd July, 2006, 08:54 PM
Member/Contributor/Resident Crystal Ball
 
Join Date: March 2004
Posts: 7,451

It will only make things slower. JEDEC timings are 7 and 5, i think, so anything close to this or faster is best, but at high HTT's, not all AMD chips can handle this, and Opteron's tend to like tighter timings even less than thier A64 counterparts.
__________________
Digg this Post!Add Post to del.icio.usBookmark Post in TechnoratiFurl this Post!
Reply With Quote
  #15 (permalink)  
Old 2nd July, 2006, 09:39 PM
gedon's Avatar
Member
 
Join Date: November 2004
Posts: 446

8 and 5 works best for tccd
2.85GHz is possible, but memory doesn't like 285MHz at 10 and 6
__________________
Digg this Post!Add Post to del.icio.usBookmark Post in TechnoratiFurl this Post!
Reply With Quote
  #16 (permalink)  
Old 9th July, 2006, 12:47 AM
gedon's Avatar
Member
 
Join Date: November 2004
Posts: 446

system isn't stable with Max Async Lateny 10, Read Preamble 6

I don't know why, but...

Max Async Lateny 8
Read Preamble 5
LDT multi 4x
htt 312MHz
cpu multi 9x
memdivider 133 -> 200MHz
vcore 1.520V

3h 20min primestable, 32M super pi stable

everything I have to change to get from 2.65GHz to 2.8GHz is htt (of course^^) and memdivider.
I thought, I tried out a 4x LDT multi last week
3x is unstable and 4x is stable at 312MHz
but, there is a new problem: mem works great at 200MHz with a 133 divider, 180 divider -> 280MHz -> totaly unstable.
I can get it super pi stable at 166/255MHz and then I get prime errors after 10sec what's going on inside my cpu?
__________________
Digg this Post!Add Post to del.icio.usBookmark Post in TechnoratiFurl this Post!
Reply With Quote
Reply



Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)
 
Thread Tools
Rate This Thread
Rate This Thread:

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is Off
HTML code is Off
Trackbacks are On
Pingbacks are On
Refbacks are On


Similar Threads
Thread Thread Starter Forum Replies Last Post
MF570sli OC Limit RenegadesRemors EPoX MotherBoards 11 1st September, 2006 04:46 PM
Speed Limit Betty Mookydooky's Just for laughs! 16 1st August, 2005 11:58 AM
FSB limit with radeon davidpowell EPoX MotherBoards 15 9th November, 2002 01:48 PM
What is the fsb limit before HDD craps out? Darkcobra Hardware Hacking 4 26th February, 2002 07:31 PM
Time limit gone? Walter:-) ThunderRd's AOA FOLDING@HOME Team 6 6th November, 2001 01:13 PM


All times are GMT +1. The time now is 12:50 PM.


Copyright ©2001 - 2010, AOA Forums
Don't Click Here Don't Click Here Either

Search Engine Friendly URLs by vBSEO 3.3.0