Well I got my replacement CPU yesterday. They sent me a whole new retail package so now I have two AMD heatsinks. Maybe I can get rich selling them on ebay! Oh wait... Anyway, I installed it and started up slowly. I ran the AMD "stability test" for an hour which it passed with flying colors. But seti was still too much for it. It took a whole 20 minutes but eventually it rebooted again. I did notice that my office was a little bit cooler than usual so I ran a test over night. I opened my window all the way and put a fan in front of it to blow cold air directly at the front of my computer. 5 hours later it was still happily crunching away. I'm doing the same thing right now. If it survives for the rest of the evening then I think I can pretty definitively say that the Phenom has some kind of overheating problem.

Note that I do not keep my apartment overly warm. It is usually around 70F (21C). The temperature sensors in the cores top out at ~50C which is warm but nowhere near critical and certainly reasonable considering I'm using the stock AMD heatsink (straight aluminum block). However the temperature sensor labeled "TMPIN2" does go up to the mid 60s which is kind of worrysome. I suspect that somewhere in the middle of the CPU (where the memory controller is) the temperature is spiking up into the 70s or higher, which is what is probably causing the reboots. My next step will be to get a better heatsink and see if that "fixes" it although I really think a CPU should be able to run without overheating with a stock heatsink unless you are overclocking.

My old 4400+ came with a good stock heatsink. It has a copper base and 4 heatpipes. That chip doesn't get above about 42C under full load. I was hoping that they were still shipping these heatsinks with the Phenoms which is why I didn't buy a better one from the start. Live and learn I guess.

---

A member of The Knights Who Say NI!
For rankings, history graphs and more, check out:
My BOINC stats site

If it survives for the rest of the evening then I think I can pretty definitively say that the Phenom has some kind of overheating problem.

As an long-term microprocessor design, manufacture, test, and reliability guy, I'd re-phrase that a little. Substantially all microprocessors run slower the hotter they get. If heating within the normal range takes them from working to not working, then, usually, the right name for their problem is they were not tested tight enough to satisfactorily guarantee operation over the expected working conditions.

Sometimes this happens, especially on new designs, because there is a rogue speed path not covered well in the test that a particular application exposes. Conservative practice is to expect this, and provide extra test margin early in product history. Needless to say, this conservatism is not popular with some important parties.

Sometimes this happens because a sub-critical defect in one specific sample creates a special case situation, just like the first, but, unlike the first, completely unreasonable to expect the manufacturer to find and cover. These are just bad luck, and everybody has them. But if you get the same type of problem with successive samples, that is not the issue--so exclude it here.

Sometimes this happens because optimists, stretched hard to meet desired specs, test with less margin than is really needed. This still does not matter unless the distribution of actual part capability contains a substantial fraction not really meeting the desired spec, but passing somewhat slower. I strongly suspect that is the situation here--that the current Phenom design, as currently manufactured, has a poor distribution relative to the currently desired specs, and that "someone" did not put in enough test margin to assure that users such as BOINC SETI users would have a reasonable assurance of success.

My next step will be to get a better heatsink and see if that "fixes" it although I really think a CPU should be able to run without overheating with a stock heatsink unless you are overclocking.

A reasonable next step, and, obviously, I agree with your expectation.

Or, assuming your motherboard allows, you could slow it down a little. For any given part, voltage, temperature, and speed are just a pretty simple tradeoff.

Of course, there can be special circuit/design issues that given an unusally steep speed, or even voltage range, deterioration with temperature. But 9 times out of 10, when user or press reports of "temperature problems" with a newish part come out, the correct description is that someone scrimped on test margin. Cyrix used to be infamous for it, just to name one old name--there have been lots, and lots, of cases.

---

Well I got my replacement CPU yesterday. They sent me a whole new retail package so now I have two AMD heatsinks. Maybe I can get rich selling them on ebay! Oh wait... Anyway, I installed it and started up slowly. I ran the AMD "stability test" for an hour which it passed with flying colors. But seti was still too much for it. It took a whole 20 minutes but eventually it rebooted again. I did notice that my office was a little bit cooler than usual so I ran a test over night. I opened my window all the way and put a fan in front of it to blow cold air directly at the front of my computer. 5 hours later it was still happily crunching away. I'm doing the same thing right now. If it survives for the rest of the evening then I think I can pretty definitively say that the Phenom has some kind of overheating problem.

Note that I do not keep my apartment overly warm. It is usually around 70F (21C). The temperature sensors in the cores top out at ~50C which is warm but nowhere near critical and certainly reasonable considering I'm using the stock AMD heatsink (straight aluminum block). However the temperature sensor labeled "TMPIN2" does go up to the mid 60s which is kind of worrysome. I suspect that somewhere in the middle of the CPU (where the memory controller is) the temperature is spiking up into the 70s or higher, which is what is probably causing the reboots. My next step will be to get a better heatsink and see if that "fixes" it although I really think a CPU should be able to run without overheating with a stock heatsink unless you are overclocking.

My old 4400+ came with a good stock heatsink. It has a copper base and 4 heatpipes. That chip doesn't get above about 42C under full load. I was hoping that they were still shipping these heatsinks with the Phenoms which is why I didn't buy a better one from the start. Live and learn I guess.

I am running the 9600 Phenom on an Asus crosshair MB (AM2) with updated bios this KINGWIN Revolution RVT-9225 heatsink assy from Newegg for like $29,with wonderful results, 46c atm running SETI 24/7, room temp is about 68F. I have swapped over almost all my rigs to this HSF assy. my Q6600's (40C- 42C) and 6700 (44C OC'd), and a Q9450 in work now, all doing superb on temps. The heatpipe direct touch it uses seems to work exceptionaly well. Just a consideration if you do decide to replace the HSF assy. I don't work for them lol, I am just impressed is all with how well it does and the price.

Well I think I have reached the end of my investigation. With "supercooling" in place the machine remained stable for over 12 hours. I have ordered a Zalman CNPS9700 heatsink which seems to score in the top few spots in most heatsink reviews.

So to summarize for those thinking about buying a Phenom: The stock AMD heatsink provides insufficient cooling for running SETI on the Phenom. You're going to want some copper heatpipe action or if you're truly crazy, maybe some water cooling :)

---

A member of The Knights Who Say NI!
For rankings, history graphs and more, check out:
My BOINC stats site

Thank you to all about the Heatsink info I was expecting a Phenom chip with the Heatsink (9600 Retail) but instead got the OEM version (Chip only in a small plastic container). I guess what was I expecting for $169.99 including the motherboard. You do get what you pay for, I only wish I had dropped the combo into my newegg basket first and realized the discount they were giving that was not advertised which was a much better deal since theirs' (Newegg's) was the retail version of the same product and the Retail version instead of OEM for about the exact same price. Now after paying for the shipping, two weeks of waiting for product and to be contacted back after a shipping DISASTER from DHL where my package looked like a hockey puck and left on my front step with no notification, the Heatsink, and everything else I need for the build, I could have had this machine built and been crunching last week!!!! Justy wanting to show my thanx again for all of the helpful info on the subject matter.

Sorry to be bitchin about my problems with FRY's Electronics and DHL's standard shipping procedures...

I have been following this thread, and recognize some of the problems peoble have seen.

Most of my problems disapeared though updating my bios to the latest version for my board (ASUS M3A32MVP).

With the newest BIOS, I was able to run my Phenom 9600 BE at stock speeds, with the stock heatsink 24/7 for days on end without problems. With the bios that came with the board, this was impossible. Overclocking was out of the question though.

Since I only ran with fan-cooling until I got myself a new water-cooling head for the CPU, I only ran with the stock cooling for a couple of weeks, but it was rock solid. I ran with the TLB fix disabled, since all reports say its not needed for desktop work.

With my watercooling setup, the temperatures have dropped 15 degrees c, and I'm able to overclock a _little_ bit.

I have overclocked the northbridge of the chip (mem controller and L3 cache (and more)) to 2.2GHz from 1.8GHz. But have only been able to overclock the cores to 2.424GHZ from 2.3GHz.

Not very impressive. I think one of the cores are holding me back, and will begin experimenting with individual clockspeeds for each core. Should be quite fun and challenging....

---

Since I only ran with fan-cooling until I got myself a new water-cooling head for the CPU, I only ran with the stock cooling for a couple of weeks, but it was rock solid. I ran with the TLB fix disabled, since all reports say its not needed for desktop work.

You are right about that unless running a server with at least two other machines running different OS accessing the server at the same time the TLB will not affect you. Or at least that was my understanding of it after reading a very long drawn out 12 - 15 page explanation, so I summed up in to one nice sentence!

With my watercooling setup, the temperatures have dropped 15 degrees c, and I'm able to overclock a _little_ bit.

I have overclocked the northbridge of the chip (mem controller and L3 cache (and more)) to 2.2GHz from 1.8GHz. But have only been able to overclock the cores to 2.424GHZ from 2.3GHz.

Not very impressive. I think one of the cores are holding me back, and will begin experimenting with individual clockspeeds for each core. Should be quite fun and challenging....

I have done some very intense research on the topic and here is a good place to start I will have to locate the other sites that have more info on it but thought this link might be of some use to you. Since you feel as if one of your cores is holding you back there does appear to be a general consensus that coer2 or the third core has issues with the overclocking and can really hold the rest back severely.

http://forums.amd.com/forum/messageview.cfm?catid=319&threadid=90533&enterthread=y&STARTPAGE=1

I know I have seen some better sites and I can send them to you directly when I run back across them, I just need to go back through my browsing history over the past week. This was just the first one that I ran back across and my IE crashed. Not surprising on this machine if you read some of my other posts... lol Thanx eMachines!!!! Good luck with the Overclocking and let me know how it is going.

Thanks for the reply rq2000.

I have read a lot about the issue of the third core (core2) holding the Overclocks back.

That does not seem to be the case on my CPU. If I clock the first core (Core0) to anything above 2424MHz, I get almost instant crashes running boinc in background. Even stopping boinc and only surfing and doing normal work crashes the system.

Keeping core0 at 2424, and setting cores 1,2 and 3 to 2525MHz, has reveiled no crashes until now. The system is completely stable at 2424 core0 + 2525 core1,2,3.

I will keep it at this setting for a few days, and then try higher settings. Wouldn't be surprised to find different limits for each core.

But in my case Core0 seems to be the "bad" core.

---

I have been following this thread, and recognize some of the problems peoble have seen.

Most of my problems disapeared though updating my bios to the latest version for my board (ASUS M3A32MVP).

With the newest BIOS, I was able to run my Phenom 9600 BE at stock speeds, with the stock heatsink 24/7 for days on end without problems. With the bios that came with the board, this was impossible. Overclocking was out of the question though.

Since I only ran with fan-cooling until I got myself a new water-cooling head for the CPU, I only ran with the stock cooling for a couple of weeks, but it was rock solid. I ran with the TLB fix disabled, since all reports say its not needed for desktop work.

With my watercooling setup, the temperatures have dropped 15 degrees c, and I'm able to overclock a _little_ bit.

I have overclocked the northbridge of the chip (mem controller and L3 cache (and more)) to 2.2GHz from 1.8GHz. But have only been able to overclock the cores to 2.424GHZ from 2.3GHz.

Not very impressive. I think one of the cores are holding me back, and will begin experimenting with individual clockspeeds for each core. Should be quite fun and challenging....

After studying comments on this thread I must now add my findings.
I use an ASUS M3A (AMD 770 Chipset) with the latest bios from ASUS.
Using a Phenom Black 9600 and stock cooler, along with Corsair Twin-X DDR-2 800 Ram, I am able to run a stable 2700MHz on all 4-Cores and get 36.5 degC.
System has been stable for 72hours now.
Ram set to 215MHz and CPU Multi to 12.5 with core voltage at 1.27v
Hope this helps someone.

I have been following this thread, and recognize some of the problems peoble have seen.

Most of my problems disapeared though updating my bios to the latest version for my board (ASUS M3A32MVP).

With the newest BIOS, I was able to run my Phenom 9600 BE at stock speeds, with the stock heatsink 24/7 for days on end without problems. With the bios that came with the board, this was impossible. Overclocking was out of the question though.

Since I only ran with fan-cooling until I got myself a new water-cooling head for the CPU, I only ran with the stock cooling for a couple of weeks, but it was rock solid. I ran with the TLB fix disabled, since all reports say its not needed for desktop work.

With my watercooling setup, the temperatures have dropped 15 degrees c, and I'm able to overclock a _little_ bit.

I have overclocked the northbridge of the chip (mem controller and L3 cache (and more)) to 2.2GHz from 1.8GHz. But have only been able to overclock the cores to 2.424GHZ from 2.3GHz.

Not very impressive. I think one of the cores are holding me back, and will begin experimenting with individual clockspeeds for each core. Should be quite fun and challenging....

After studying comments on this thread I must now add my findings.
I use an ASUS M3A (AMD 770 Chipset) with the latest bios from ASUS.
Using a Phenom Black 9600 and stock cooler, along with Corsair Twin-X DDR-2 800 Ram, I am able to run a stable 2700MHz on all 4-Cores and get 36.5 degC.
System has been stable for 72hours now.
Ram set to 215MHz and CPU Multi to 12.5 with core voltage at 1.27v
Hope this helps someone.

I can not believe that you are getting 36.5 degres. That is astonishing...

I can not believe that you are getting 36.5 degres. That is astonishing...[/quote]

Current Room Temp is 15 degC all 4-cores are at 35/36.5 degC.
If it's any help, I build PC's here in the shop and the case is a Coolermaster RC-690 (Dominator). This case uses 3x 120mm fans one of which is above the cpu extracting any hot air directly to the outside and the psu is at the base of the case.

Well dispite me having gotten myself a bad 9600BE, I'm still working with the overclocks. 2424MHz is the limit for Core0, unless I use extreme voltages like 1,5V, which I'm not comfortable running at for longer periods of time.

But in the meantime, I have gotten the three other cores to 2626MHz, without any problems. Going to try faster speeds later today.

My only problem is getting these settings to load at startup. AMD Overdrive supports doing this, but when I enable this setting, either nothing happens, or the system freezes during bootup :-(

Setting it manually, works, but I would much rather the BIOS would let you set the cores individually. But thats probably to much to hope for.

Another problem with AMD overdrive is, that sometimes when I shut it down, it resets the cores to their bootup speed, making me have to enter again, to set it up once more. Very annoying...

The temps q3xkr mention sound astonishing to me to. Im running 38-41 degrees celcius, and this is with (very quiet) watercooling. I could get lower, but that would make my system much more noisy, and I prefer the silence :-)

---

The temps q3xkr mention sound astonishing to me to. Im running 38-41 degrees celcius, and this is with (very quiet) watercooling. I could get lower, but that would make my system much more noisy, and I prefer the silence :-)

I have no way of knowing, but the cpu I have is a trade sample from AMD issued to system builders / partners to assist with promoting the Phenom.
It could be (but no proof) that the cpu I have is "Hand-Picked".
I have found using Crunch 3R client that at 2700MHz the Phenom locks up occaisionally. However at 2620MHz all stable for over 48hrs.
I can achieve 2600MHz by either the multiplier on AMD Overdrive or tweaking the ram to 210MHz.

---

The temps q3xkr mention sound astonishing to me to. Im running 38-41 degrees celcius, and this is with (very quiet) watercooling. I could get lower, but that would make my system much more noisy, and I prefer the silence :-)

I have no way of knowing, but the cpu I have is a trade sample from AMD issued to system builders / partners to assist with promoting the Phenom.
It could be (but no proof) that the cpu I have is "Hand-Picked".
I have found using Crunch 3R client that at 2700MHz the Phenom locks up occaisionally. However at 2620MHz all stable for over 48hrs.
I can achieve 2600MHz by either the multiplier on AMD Overdrive or tweaking the ram to 210MHz.

Just out of curiosity, did you remove the heatsink and add some different thermal grease? I have seen many people saying even running stock speeds with BOINC going with the cooler provided they have to back it down for fear of frying the chip(stock cooler wasn't adequate for it)...I can't imagine how with yours being overclocked and having a stock cooler even with freezing room temps isn't running hotter than the "average bear". Not trying to discount your claims. Just not sure what you are using to determine the core temp while it is running at these speeds for so long and if it might, perhaps be a little flawed or possible sensor mixup. Not that mine is finished being built yet I need a decent heat sink and have yet to decide which will be right for me without hurting my pocketbook more than it has already. Don't get me wrong we would all LOVE to have hand picked versions of our own chip that would OUTPERFORM everything else available IT MUST BE NICE...
Just another thing I read recently ASUS doesn't have a board capable of supporting the newest 9850 or any 125 chipset and have not realesed any info on when an updated Bios will be available to allow these chips to work in ANY of their motherboards. Some of their competitors have some under $100 that will run these same chips...Just some info to think about before anyone decides to UPGRADE to the newest Phenoms with an ASUS motherboard

I think I heard that the black edition comes with a decent heatsink like my x2 4400+ came with. I could even see that one holding up to some overclocking. It has a copper base, 4 heatpipes and aluminum fins.

However the "normal" Phenom that I have comes with a chunk of aluminum that can't even stand up to stock speeds. I have positively confirmed that this was my problem. With my new Zalman 9700 the CPU runs a full 20C cooler (!!) under load and basically idles right at ambient. It even looked like it was idling below ambient (15C) but I'm guessing there was a cold draft coming in from the window at the time. Everything is running at full (stock) speed and I haven't had a reboot since the new heatsink was installed.

Note: installing that beast of a heatsink was no simple matter! I just uploaded a picture of it after installation to my profile. It really dominates the case now.

---

A member of The Knights Who Say NI!
For rankings, history graphs and more, check out:
My BOINC stats site

I think I heard that the black edition comes with a decent heatsink like my x2 4400+ came with. I could even see that one holding up to some overclocking. It has a copper base, 4 heatpipes and aluminum fins.

However the "normal" Phenom that I have comes with a chunk of aluminum that can't even stand up to stock speeds. I have positively confirmed that this was my problem. With my new Zalman 9700 the CPU runs a full 20C cooler (!!) under load and basically idles right at ambient. It even looked like it was idling below ambient (15C) but I'm guessing there was a cold draft coming in from the window at the time. Everything is running at full (stock) speed and I haven't had a reboot since the new heatsink was installed.

Note: installing that beast of a heatsink was no simple matter! I just uploaded a picture of it after installation to my profile. It really dominates the case now.

My 9600 BE's came with the same heatsink as my Athlon X2's did, a copper one with heat pipes.

---

I have a couple of comments...

The heatsink that came with my 9600BE, was a aluminium with copper base and copper heat pipes. It ran my Phenom reliably at stock speed, in my not very ventilated, and hot case.

Regarding ASUS. I run the ASUS M3A32-MVP deluxe. I'm quite certain this board supports _all_ Phenoms that are out, with the latest 1002 bios. I would be surprised if it didn't, as it's a 8+2 phase regulated board, build with overclocking in mind.

Lastly, I have had more experience with OC of my Phenom. In my last post I thought it was table at 1x2424 an 3x2626. That wasn't true. Core2 (CPU3), would die after many hours at this speed, so its back to 2525, where its stable. I'm trying to get Core1 and Core3 to 2727MHz, but haven't gotten complete stability yet. Core3 seems most villing to get up there.

I would find it a little funny to run Core0 at 2424, Core2 at 2525, Core1 at 2626, and Core3 at 2727 :-) Allthough it would be much more fun to run all of them at 3500Mhz....

Anyway, this CPU will have to last, until I can get a 45nm phenom with 6Mb L3, and about a 3Ghz (3GHz+) official speed. I hope that day is not too far away....

---

I also fired up my first Phenom, the brand new B3 Stepping X4 9550.

Reading about "not galactic performance", I was very pleasantly surprised to see each 2200MHz Phenom X4 core outperforming my fastest 2800MHz Athlon64 X2 core (Crunch3r V2.4V 64bit Linux optimized Science Application).

AMD seems up to specs with the manufacturing process as well again, my Phenom X4 9550 operates at default 1.200Vcore, but it maintained Prime-Stable all the way downto 1.075Vcore (something I definitely didn't expect, after what I heard about the 9500/9600 B2 Stepping)

Now it's working at 1.100Vcore, stock 2200MHz and the entire System (max. power-saving equipped) consumes anywhere from 94W (4x BOINC/MalariaControl) to 110W (4x 8MB Prime95 max.CPU-load Instances).

Bottom line : It's cheap and it rocks

---

FalconFly:

Glad to see your happy with your Phenom.

In comparison to my old Athlon64, my Phenom is substantially faster at the same tasks, even if they are single tasked. I think the reputation it has gotten as a bad performing processor is a bit undeserved. Its not as good as Intels Core series, but its still a big step forward for AMD's lineup.

I also find some of the reviews quite misleading. One review had a Phenom 9600 much like mine, and on the same motherboard, perform terrible in a Winrars compression benchmark. It benched at about one fouth the speed of a Core Quad. When running the exact same test on my phenom, it shows numbers that were about 5-10% lower than those published for Core Quad. Phenom _is_ not as fast, but its not as bad as its reputation.

I will just add, that my Phenom is now running 1x2424, 1x2525, 1x2626 and 1x2727 Mhz on Core0, Core2, Core1 and Core3, for almost 24h, and the nothbridge of the chip (L3 cache and more), is also overclocked from 1800Mhz, to 2222MHz (each 200Mhz here gives about 2-3% higher performance). To do this I had to raise core-voltage to 1,35V, while the northbridge voltage stayed at 1,25V.

Being able to adjust all the cores, and the cache/northbridge speeds individually, is actually the most impressive thing about this CPU IMHO.
To bad AMD is so far behind in process technology / circuit layout, that they are having problems ramping up the clocks, but I hope that the next respin off the Phenom (to 45nm with 3x larger cache) will be a more competitive product. Hope they have the process and circuits of the chip optimized, for 3GHz++ at that time.

---

When you see a review that has the Phenom do poorly in Winrar, it's because they've enabled the BIOS "fix" for the L3 cache race condition errata. However now that the B3 step is out (9x50 models) the paranoid no longer need to have this enabled and the Winrar benchmark is much, much better.

---

"Life is just nature's way of keeping meat fresh." - The Doctor