Phenom X4 9850 BE

[AJ.]

Just wanted to make a new thread to document my oc'ing. So far I'm at 2.6Ghz. I just upped my multiplier to 13x. I read a couple reviews about this chip and it doesn't seem to like increasing the FSB much.

http://img155.imageshack.us/img155/756/screenhunter01jul230558ce5.gif

Anyway, it seems pretty stable. I'm running some stress tests in AMD overdrive to see how stable it is. So far so good. The only thing that's worrying is the temps. When all cores are maxxed out my temps range from 49.5-51C (mainly 50C). I know these chips do run hot, though. At idle my temps stay between 40-41c.

[PP Mguire]

Your fine thats not exactly hot. 2.6@1.28 is pretty decent actualy. You might be able to pull off 3.0ghz pretty fine that way maybe.

[AJ.]

I'm already at 2.74Ghz stable. Temps are within the same range.

http://img186.imageshack.us/img186/8099/screenhunter01jul230629vh4.gif

[SonicWRX]

Good work AJ! Keep us up-to-date on this one I will be getting one of the AMDx4 chips soon. I'd like to see what you can get out of it.

Mind telling us how you are cooling it?

[AJ.]

Thermaltake Mini Typhoon. It's not the best cooler, but it does get the job done. ;-)

The one that came with the processor is quite impressive, though. It had copper heat pipes and everything. I might just pop it into my server.

Btw, 2.8Ghz stable.

[AJ.]

Well, I think this is all I'm going to do for now. I want to balance out performance and longevity. 2.8Ghz is plenty fast. If you really wanted to, you could easily push this baby to 3Ghz. I'd go as far to say 3.1-3.2 if you have the right gear and cooling.

It's an excellent chip. I've had experience with a Q6600 and it performs exactly the same. While the Q6600 oc'd much faster and easier, the AMD wasn't exactly hard to OC. I'd give it 9/10. :)

http://img293.imageshack.us/img293/9533/screenhunter01jul230729xz5.gif

http://img244.imageshack.us/img244/5804/screenhunter02jul230729xj4.gif

[jellyrole]

Great overclock on that chip! Glad you like it.

[Victor]

That seems to be what I was able to get out of the x3 8750 as well. Phenom does not seem to be able to break 3GHz that easy but newer mobo seems to be able ot push it a bit farther than old board. I do not have the 140W version to test it but I think those OC slightly better.

[AJ.]

I think you're right Victor.

I know I could easily go for 3Ghz. But my board is brand new and was released and designed specifically for the launch of the Phenom. So I'd imagine it would OC better on it.

I do not have the 140W version to test it but I think those OC slightly better.

The 9950 isn't worth it IMO. It uses too much power and you don't gain much over getting the 9850. But you're right about it OC'ing better.

[Xero (1)ne]

Yah I think the multiplier is better for overclocking these than teh FSB. I got mine up to 2.8 but because of horrible airflow issues in my new case it got way too hot for comfort so now I'm at stock.

[MtDew]

I have got my Phenom X4 9600 BE running stable at 2.6GHz (13x200) right now. I was able to push it to 2.7GHz (13.5x200) at stock voltage, but it would start to flake about 11 hours into running prime. I am getting temps about 48c at full load at 2.6GHz. I am getting ready to leave on vacation, so I am going to back it down to the stock clock when I get back, then I will work to get this up above 2.8GHz.

Good luck pushing yours to 3.0GHz.

[FunkZ]

There is nothing wrong with running higher bus speeds with these chips, you just have to remember that overclocking by bus not only affects the processor speed but also memory, northbridge and hypertransport speeds, and you may need to adjust memory ratio and NB/HT multipliers to compensate. Here's a link to a Phenom overclocking guide I put together: http://www.ocia.net/articles/k10overclock/page1.shtml

I'm running 3.0 Ghz with just a small bump in voltage and good cooling, temps are low to mid 40's at full load. I ran 250x12 for awhile when I was using 1066 memory. I've since switched back to my Titaniums and had to lower the bus to reduce memory speed, currently 230x13, but at lower timings. NB/HT Multi is at 9x.

BTW your memory controller is ganged, meaning it's using 1x128bit instead of 2x64bit.

[AJ.]

Very nice article Funk. I'll give my CPU a shot at 3Ghz with some of the tips you provided. :icon_tiphat:

[Victor]

yes, that's a very good read for thos want to oc their cpu. Though I think it works better with amd based chipset. With the upcoming amd chipset that promises to give additional oc potential, I think OC 3GHz is within the reach. Still, I am waiting for the 45nm version of the chip as I think those are the one that can easily reach 3ghz or higher.

[AJ.]

So far I'm at 2.9Ghz and I'm testing stability. My motherboard didn't like the route Funk took. I tried adjusting them slowly to no avail. I couldn't even get to the welcome screen without the BSOD. I'm probably going to hit 3Ghz by increasing the voltage on my CPU and increasing the multiplier. So far things are looking good. :)

[AJ.]

Hmm....It looks like I'm going to need 1.45+ volts to keep it stable at 2.9Ghz. That's a little too much for my taste lol. I'm probably better off leaving it at 2.8Ghz and reducing my memory latency as much as possible. I'll pass on the 3Ghz challenge. I want my CPU to last. lol :lol:

Btw Funk, I get better performance in ganged mode than I do in unganged according to Everest.

[FunkZ]

Glad you liked the article, and congrats on reaching 2.8 ghz. As you pointed out, memory bandwidth benchmarks like those in Everest and SiSoft Sandra will give you higher scores using ganged mode. Those tests will also favor higher speed memory over lower latency. Other types of "real world" tests however can do better with unganged memory running lower latencies. It depends on the benchmark.

[RiverRicer]

Ya.....2.8ghz was about all I could get out of my 9850BE/Jetway HA04 Extreme with any sort of stability. I just installed the 9850 on a new Gigabyte MA790GP-DS4H (790GX/SB750) and here's the stable result @1.36vcore:

http://valid.x86-secret.com/show_oc.php?id=400966

No more freezing and definitely faster drive read/writes.

[PP Mguire]

Nice 3.2. Seems they are getting better at OCing maybe.

[RiverRicer]

The later B3 silicon seems to be better than the chip I've got - a 0809......probably off the first wafer!:dielaffin:
I'm seeing 3.4-3.5 from some of the users with SB750 motherboards and later silicon over at Xtremesystems.
There's also a thread on the Deneb (45nm) with a 4.0/1.475v overclock on air. It's a C0 stepping engineering sample.
AMD really got it's ass handed to it this past year, but I think they're finally going to turn the corner.............first with the suprising 48xx GPUs, now with an excellent M/B chipset......just need a more competitive CPU, which is apparently coming!
BTW, this o/c has trimmed around 1 minute off each tick of Project 2665.:biggrin:

[AJ.]

Squeezed another 100Mhz in on stock volts. I'm now running @ 2.92Ghz (14 x 208 ) @ 1.28V. Idle 37c and 48C load. I'm still running some stability tests. So far it's looking good.

Will keep you guys updated. I want to see if I could get more out of this baby at stock volts. :)

http://img177.imageshack.us/img177/6277/screenhunter02aug241317cr1.gif

['Amp!']

what is the overall system wattage like?

Full system specs also please.

[AJ.]

Thermaltake Toughpower 650W PSU
Gigabyte MA-790FX DS5 (BIOS version F6)
2GB Kingston HyperX PC28500 1066@ 1111.2Mhz
X4 9850 BE (14 x 208.35) 2916.9Mhz
ATI X1300 Pro PCI-e (temporary lol)
Segate Barracuda 320GB HDD
WD Caviar 250GB HDD
Zalman CNPS 9500
Silverstone FP52

I'm not too sure on how my overall wattage is looking like. With my X1300 in there for now, I'm probably anywhere between 150-200W.