Jump to content

Abducted

Members
  • Content Count

    51
  • Joined

  • Last visited

About Abducted

  • Rank
    Member

Contact Methods

  1. I'm so happy AlexS asked this question and Sailindawg was here to answer I've been having a lot of problems with the 'memory controller' on this new X2 4400+ 89w. When running prime95 CPU stress test on both cores there is no problem, when using blend test core0 always fails way before core1. which is kinda weird because the difference between just CPU stress test and blend mode is the stress on the memory controller and memory and since there isn't a seperate memory controller for each core i didn't quite get this problem i was having. And eventhough this LDT voltage still doesn't really explain it it does help me. Not sure how much yet, but did a quick test at 290HTT with the memory at DDR580 and with 1.2v LDT prime95 failed after a little over 2min. Now with 1.3v LDT it's been going 40min so far. Of course if this is all i get out of it, it's nice for me, but nothing special. But if with this 1.3v LDT i can make sure core0 overclocks the same as core1 i'll be very happy. As for some of the reaction u were getting in that thread u linked to Sailingdawg. Saying LDT doesn't help them because there system is perfectly stable the way it is of course isn't a good arguement. If i run my system at 10Mhz above stock it'll be perfectly stable with 1.2v LDT too. That doesn't mean if i push my system that at some point the LDT could be holding me back. Yes they might be stable at, i dont know, maybe 9x330Mhz. That still doesn't rule out the fact that they might be able to push it to 9x331Mhz if they increase. And of course we with any voltage setting, changing it only has effect if the part that's holding you back. If your memory has a limit of 270Mhz at 2.8v then increasing the Vcore isn't going to make your memory overclock any higher. Most of the guys are overclocking opterons and i think those come with better memory controllers (and i guess better LDT busses) so maybe in their case the LDT voltage isn't going to have any effect because because it isn't what's holding them back. Anyway long story short. i didn't know the LDT voltage was for a bus on the CPU die, i figured it was another voltage for the nF4 chip. And it looks like it'll be helping my system overclock. How much help i'll report back later. EDIT: Oke by changing my LDT voltage from 1.2v to 1.5v i went from unstable at 290Mhz to stable at 292Mhz, so lets just say that 3Mhz HTT giving me a 27Mhz higher clockrate as the same Vcore. I think that's a nice result from just adjusting my LDT voltage.
  2. Another 89W 4400+ here from Europe. ADV4400DAA6CD CCBWE 0613BPMW. [email protected] Looks like it isn't doing that bad when i compare it with the results in this thread. But i did expect more from it really. The memory controller on this thing sucks big time. It requires a lot of finetuning and it can even make the system totally unstable at stock with the wrong settings (yet settings my memory can handle for sure) I can have it to 2700Mhz with 1.4v and that's not really bad, but like mentioned before i also hit a wall at 2.7Ghz. These CPUs use the same Vcore as the 110w version if i'm correct, so they did something else to lower the TDP and i think it has something to do with the memory controller or of course it could just me mine that has a crappy controller. Before i bought the CPU i checked amdcompare.com and it said the 89w version used a lower Vcore and just like the 1700+ JIUHB DLT3C 1.5v instead of 1.6v, the Duron Applebred at 1.5v instead of 1.6 the Athlon XP Mobiles at 1.45v i expected it to be a great overclocker. Turned out the Vcore isn't lower and when i checked the website again they changed the Vcore rating for the ADV version!!! It does run cool. I'm using watercooling with 2 radiators, without the IHS and with the latest Shin Etsu and with dual prime it doesn't come over 35c! It's basicly the coldest chip in my system.
  3. You should also use full multipliers, no 0.5 multies. I suggest you just leave it at 9x. I got my 4400+ last week and you can see the stepping in my sig. It's a ADV version meaning its a 89w version instead of a 110w. But on the AMD compare website the Vcore shows up the same for the 89w and the 110w version. So they must have done it in another way. Anyway, it's not a great overclocker and i have a feeling it's mostly the memory controller beeing a real biatch. It needs extreme finetuning to get it to a higher stable overclock. Here is a list of what mine does at variating voltages. 1.250v = 2442Mhz - 1.300v = 2519Mhz 1.325v = 2563Mhz 1.350v = 2607Mhz 1.375v = 2662Mhz Things start to get rough after that. I have tried any extreme voltages yet. My single core Opty did just 2350Mhz at 1.250v, but did 2900Mhz at it's stock 1.375v (1.4v in BIOS). I assume this is a normal difference between single and dualcores?
  4. When your CPU temp is 64c or even 50c and the heatsink is properly seated the heatsink should feel really hot near the base. If it doesn't then either it's not seated correctly OR there is something wrong with your temp sensor.
  5. Normally a 25c limit would make sure that the fan stays at 100% all the time. Your cooling it too good that's your problem! Maybe you should get the latest BIOS from the DFI website. You can 'adjust' the CPU temp in that BIOS. Just set it a little higher to that the CPU temp doesn't get below 25c. I guess that should work.
  6. Maybe it's just a DNS server problem, meaning that some domainnames do get resolved and others dont. Since ure using a router i assume the DNS server in your TCP/IP config is the IP address of the router. Try connecting to the DNS server directly by entering the DNS server IP of your ISP. You can probably find out what IP that is by checking the config page of your router or by checking your ISP's website. If you can't connect to that website just use (one of the) the following DNS server(s) to check if that fixes anything: 62.179.104.196 or 212.142.28.69
  7. Hey folks! I hope you can take the time to read the full post. I've been running an Opteron 146 at 3Ghz (1.475 in BIOS) for a long time with the stuff in my sig. The memory can do about [email protected], [email protected] and [email protected] 2 days ago i got an ADV X2 4400+ (89w) and installed it. I did not reinstall windows (Win 2003 SP1 32bit) just so you know. I did get the AMD Dual-core optimizer and the patch for WinXP (wasn't sure if it applied to Win2003 too) At first i used stock CPU settings (11x200) and memory settings op 2.5-3-3-7 (1:1) and prime95 even failed at those settings after a while, 2.5-4-3-7 did worse and 2-2-2-6 did best of the 3 settings but still unstable. So basicly i thought the CPU was just messed up cause i just couldn't get it stable using the 11x multiplier at stock speeds. But when i tried 10x220 everything was stable (for as long as i tested). Then i started to increase the HTT and it got unstable at +/- 10x240Mhz then i changed the memory timings to 2-2-2-6 (166Mhz divider) i could push it to 10x245Mhz. Increasing the voltage didn't help so i checked to see if i could lower them and i could lower them from 1.35v to 1.3v for the CPU and 2.7v to 2.6v for the memory (to keep the ratio 1:2). So i guess it wasn't the voltage holding me back. Now i cant seem to get it any higher and 2450Mhz is a crappy overclock for this CPU as far as i can tell. I have a feeling i either overlooked something or something is 'broken'. So far memory setting changes have had the most influence so i would think it's a memory controller issue. But then why can i run 10x240=2400Mhz with the memory at 200Mhz using the 166Mhz divider and not 11x200 and the memory at 200Mhz?! It could have something to do with the mainboard, with the Opteron of course i couldn't use the 11x multiplier, but the 10x worked without any problems. It could have something to do with the memory, but it's been stable at much higher speeds with the Opteron and even now it is stable in memtest86+. It could be my windows installation, but i have also tried running Prime95 from the ultimate boot CD (UBCD) (just single core) and that produces errors about the same as what i see in windows. I was running the latest BIOS (4/06/2006) and now i'm using the modded 704-2BTA but that didn't change anything. I've tried my memory sticks in the orange and yellow slots. I've tried those errata options in the BIOS on auto and disabled, no change. Can somebody shine some light on this for me please? I've found one thread in this forum from somebody with what seems to be similar problems, but there is no solution there. Any suggestions would be welcome. Thanks! EDIT: Just so you know i can do 2350Mhz at 1.225v, my Opteron could only do 2300Mhz at that voltage. Also i can turn off the fans of my watercooler and that way my water temp goes up from 29c to 42c and even that doesn't change much.
  8. Did you find a fix for this problem yet? I just bought an ADV X2 4400+ (89w) and it's VERY unstable when using the 11x multiplier too, even at the stock 11x200. The 10x multiplier works fine. So far i've tried the latest BIOS (2006-04-06) and 704-2BTA and i noticed to difference. So before trying every single BIOS i thought i check here to see if this was a common problem. Well it's not common i guess but atleast i found this thread. Too bad there is no solution yet.
  9. Hey folks! I'm thinking of upgrading my system 'soon'. I was thinking about getting a NF4 Ultra-D with either a 3200+ Venice or a 3700+ San Diego for it's bigger L2 cache. But now i've found a store that sells S754 A64 Mobile 3000+ LowPower (35w) and those are cheaper then the 3700+, also have 1MB L2 cache and i hope will overclock even better then the San Diego. (altough i just noticed those already run at very low Vcore of 1.35/1.40v, not sure what those lowpower mobiles run at) I know the difference between S939 and S754 is Dual Channel and Single Channel. But that's supposed to be just a small performance difference (+/- 5%). Of course a good choice for the S754 is a NF3 250GB, but the new NF4X Infinity is also S754, uses PCI-E and it is a newer generation chipset. But i can't find any information about the board other then what's on the DFI website. I'm thinking of buying some Winbond UTT memory so a high Vdimm would be nice and i know the 250GB can go up to 4.0v. What can the NF4X do? What kind of HTT is average for a NF3 250GB and what would you expect from a NF4X? Thanks! p.s. if i go for the S754 combo it'll be 100euro cheaper.
  10. Why do you guys have such a problem with a bios not recognizing your mobile CPU? :confused:
  11. You can enter the MAC address in the bios or under device manager in windows. Have you tried both? I know that if you enter it in bios you'll find out if the MAC address is the long or short one. If i remeber correctly it's the short one because if you try to enter the long one it just won't go any further then 11 characters.
  12. Well since nobody even replied here i'm sure there are more DFI users who don't know about BPL versions.
  13. Well maybe BPL version isn't such a big thing with DFI. BPL is a nVidia build part of the bios file and u can 'mod' it into any bios file without changing the rest of the bios. On the ABIT board there were atleast 8 differend BPL versions that worked for it. But i know there were a few that didn't work on the NF7-S. With my Samsung PC3200 C3 and A-DATA PC4000 C3 memory older BPL versions worked better. Meaning that if i used a new bios version modded with an older BPL version it would work better. With my BH-5 it was the other way around. The newer BPL versions worked better. Like i said a difference in BPL version would have quite a big influence on my NF7-S. So i would expect the same for the DFI boards.
  14. With the ABIT NF7-S a differend BPL version had quite a big effect on how my system overclocked and the right BPL version increased 3D stability by 6Mhz FSB for me. But if i'm correct not all BPL versions work on all nForce2 Ultra 400 boards. Has anybody tested yet what BPL versions work on the DFI boards? And if so are those bios file available somewhere?
×
×
  • Create New...