Athlon ii x3 435 dead?

Page 3 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.

Josh154

Distinguished
Jan 27, 2010
85
0
18,630
Hello all, i just built myself a rig today. It was my first time building a computer. I am having some difficulties with it right now though. First here is the specs,
-MSI 785GM-E65 mobo
-AMD Athlon ii x3 435 processor
-500gb samsung 7200 rpm 16mb cache hard drive
-2gb super talent 1333mhz ram
-Raidmax 450w psu (came with case and im on a VERY tight budget)
-LG dvd burner
-Windows 7 Ultimate 64 bit

Anyways here is my problems. I can boot into windows and do some basic browsing on the internet. Whenever i go to play a game using a large amount of the CPU the system hard locks up. When it locks up i cannot move the mouse or turn off num lock or any of those options on the keyboard. When it locks up it freezes and sometimes it will go to a black screen with white lines running vertical down it or it will be at whatever task im doing and some red lines start to show up.

The processor is not overclocked in anyway. I have unlocked the 4th core on it and had it 3.5ghz and was still able to boot into windows and browse the internet. Nothing else, it is back at stock configuration now and still having problems.

I have tried running prime 95 and as soon as i start it no matter what test it locks up. Sometimes it will run for a minute or two then lock up. The max temp i have seen while running prime is 49 Celsius. The max on this chip is 73 Celsius i seen on the AMD site.

I have ran memtest 86 and my ram has passed that. Everything in the BIOS is stock configuration. Just as soon as i go to use a large CPU percentage it locks up. Should i try manually adjusting the vCore? CPU-Z shows it at a core voltage of 1.424, doesn't that seem high for a stock chip not OC'd?

I have run out of ideas. Should i just take the thing back? I bought it at a local shop for $85. Is it actually a bad chip or something i need to change?
 
Wow, as Razbery pointed out, that really goes to show the power of ACC. Changed the number of cores > Changed the CPUID String > Changed the voltage.

Wow, so you pretty much transformed this chip into another one! :O

Wow...
 
Technically, he didn't transform anything... Seems as though every chip that runs off the line is a Deneb quad.

Some AMD CPUs come off the line with flawed cores, which AMD then disables. This leads to X2 and X3 models. I would suppose this method is what leads to X2's and X3's which either don't unlock, aren't stable when unlocked, or don't have much OC headroom when they are unlocked. However, sometimes the demand for X2 and X3 models is so high, AMD's forced to take functional quads, sever a core or two, and release them as X2's and X3's. I believe these to be the ones which more successfully unlock, but not only unlock and remain stable, but also often OC quite well.

That's just my thoughts on it... I dunno if that's exactly how they do it, but it makes sense if it's even anywhere close.
 
Well, yeah. You can't change the arch. of a CPU, but you can change how the OS/rest of the hardware recognizes it, effectively "changing" the chip. (If you get what I mean).

And yes, that's correct. It's cheaper to manufacture x4s only, rather than x4s, x3s, and x2s all at once. So, during the binning process, AMD disables cores. And since they don't have equipment to test every single chip on the line, they check a few dies from the same wafer to see how everything is going, and disable accordingly.

It's a rumor that chips from the center of the wafer perform better than ones on the outside, so you probably got one from the center. 😛

It's the same story with the 5870s and the 5830s. ATi/AMD bins them accordingly, and some claim the higher title of the 5870, while lesser-destined chips find their way into the 5830 bucket. 😛

The odds of enabling cores that actually work are pretty much the same as a 4GHz overclock on an i7 920/930 as far as I've heard. They're almost in your favor, but it's still a gamble.

Cheers.
 

In this case the cache is also disabled.
 
L1 and L2 cache are part of the core (so I've been told). So, naturally, unlocking the core will unlock cache.

L3 cache is the only real architectural difference that separates the Phenom II and Athlon II.

Sorry to hijack your thread. 😀

Cheers.
 
I actually got a 425 not the 435 :) They were sold out of the 435's so i just went ahead and grabbed a 425. Just got home and had been running prime for about 7 hours i think shes stable :) Time to start doin some overclockin to this beast!
 
Hmm... I am trying overclocking right now. I bumped it up to 2.9Ghz and not stable. I think i need to up the voltage a bit. In my BIOS it just shows the voltage as auto, i cannot figure out how to change it. Anybody think i need to up the voltage a bit to get it stable? 2.9Ghz is the 435 stock speed and its voltage is higher than the 425 at 1.320 and 2.7 clock speed.
 

Well some of them are unstable at overclocking that's why they finish with 3 cores and no l3 cache.
 
Congrats on the unlock! I wish i could get mine going that speed 🙁 I can't seem to get stable at anything over 2.9Ghz. I think i need to up some voltages to get it stable.

Raz, would you mind posting a CPU-Z shot?

Also would anyone else let me know what would be quicker? Quad core with 6mb l3 cache at 2.9Ghz or triple core with no l3 cache at 3.2Ghz?
 

I would choose the unlocked quad with the unlocked L3 cache. You won't see a big difference either choice though, but the L3 cache is a bigger benefit than a 300mhz OC.
 
Sorry for the wait, Josh. Take a peek:



If you click it, it'll show everything. I'm not so sure it's entirely stable, though. It Prime95'd overnight for 5 hours just fine. But, it caused a BSOD today almost 2 hours into an identical run. I had to test it again because for some unknown reason, when unlocked, this chip makes my system think that the default PCIe clock is 108 MHz instead of a flat 100 MHz. I had to fiddle with it a freakin ton to get my PCIe clock straight, but it also added a little more "fluff" to the already slightly inflated FSB. (Default is 200, but it clocks at 200.8 - 200.9 Mhz. The added "fluff" took it to 201.6 MHz.) With the added 0.7 MHz to the FSB clock, I had to test for stability again. I have a feeling if I just take it to 1.35V or 1.375V it'll stabilize (as a Quad) at 3.222 GHz.