I attempted to run Real Temp on my ex487 and it would come back with some error about the network file. I don't have time to look into it any further today. I'll try again tomorrow and see what happens.
EX487 e5200
Categories: Server Mods and Hacking
Hi Everyone:
I just upgraded my brand new HP EX-485 WHS. I used the Intel Dual-Core E5200 (2.5GHz, 800 FSB, 2MB L2 cache).
I followed Donavon West from www.HomeServerHacks.com instructions. (Thanks Donavon).
The server is working just fine. There are no issues to report.
Here's some pics:
Thanks to everyone on this great forum and the entire WHS family.
Marcos
-HP EX-485, Intel E5200 CPU upgrade, 4-750G HDs
I attempted to run Real Temp on my ex487 and it would come back with some error about the network file. I don't have time to look into it any further today. I'll try again tomorrow and see what happens.
EX487 e5200
k, thanks Jeffla. Did you unblock the file in the file properties settings? I'm running it from the servers desktop.
Ok, so I got Real Temp working and did a few screen shots with from left to right CPU-Z - Real Temp - Core Temp. As you can see, the 1st pic shows a 100% load after 5 mins. They all correctly read a 12.5 multiplier. The second pic shows an idle state after the load test. As you can see, they all report different numbers. Below is a chart of the max and min I have ever seen with each program.
Program - Max - Min multiplier
CPU-Z - 12.5 - 6.0
Real Temp - 12.5 - 11.0
Core Temp - 12.5 - 9.5
I've done zero modification to any of these programs, I just simply installed them and executed them. I guess I have no idea which one to believe is accurate. No matter what I do, I can not get the multiplier to drop when under load in an attempt to mimic your situation. My CPU behaves as expected 100% of the time. The only odd thing is why each program has a different base (min) multiplier. If I do re-enable speed stepping 01 01, they do all correctly show the processor locked at a 6 multiplier.
http://picasaweb.google.com/lh/photo/lPueJSaINeq_RUOs30N2AXHiAUBkejRhh-Go2TfS...
http://picasaweb.google.com/lh/photo/833IOOFxuEH0EaI9nfuvVHHiAUBkejRhh-Go2TfS...
EX487 e5200
Jeffla, thanks for the test numbers!! Can you do one more thing... Actually two things.
First, what version of HP Software are you running? I'm running the updated version 3.
Second, in the Real Temp folder there is a program called i7Turbo.exe . Can you run it and check/uncheck the C1E box and watch multiplier behavior?
Thanks!
I'm running HP's 2.5 software.
i7turbo.exe does the following. With C1E enabled, it reports ~11-12.5 multiplier, with C1E disabled, it reports a constant 12.5 multiplier. With C1E enabled, under load it stays at 12.5 until the load ends, then it fluctuates between 10.841(or something like that) and 11 at idle.
EX487 e5200
Thanks again Jeffla! It looks like something weird is going on with network transfers because every other behavior is exactly the same... Maybe I'll try to install Windows 7 on it just to eliminate the OS variable. Thanks!
I've read the article on upgrading the EX48* using an E5200 (as well as many of the posts on this forum) but can't find the requisite M0 stepping version anywhere. All I can find is the newer R0. Anyone know where I can get an M0? I understand the R0 stepping version requires a firmware upgrade which I was disinclined to perform without a KVM and did not want to spend the money on it. Is that not a big deal? I don't see anyone discussing which version they used. Most posts just say E5200 and nothing else.
Any feedback on this would be greatly appreciated.
Sorry, I have the R0 too and I definitely would have been in a bind without the KVM.
Even with the M0 SLAY7 e5200, you're still going to need access to the bios to disable speed stepping. Otherwise your processor will be locked at a CPU multiplier of 6 instead of 12.5. You don't have to disable speedstepping, but then you'll never get full speed out of your CPU.
I just flashed the changes to my bios and was done with it. I bought the MO SLAY7 e5200 thinking I wouldn't have to make any modifications to my bios, but I quickly discovered that disabling speedstepping was necessary for the e5200 to achieve anything over 1194mhz (6 multiplier). Now with the changes in place, my e5200 M0 SLAY7 works perfectly and will, when needed ramp upto 2487mhz (2.5ghz) 12.5 multiplier.
So don't search for a SLAY7 M0 thinking you can simply drop it in and achieve the cpu's full potential (2.5ghz). It is true that the M0 SLAY7 is a drop in upgrade requiring no bios modification to allow it to boot and function, but of course, it will only function at 1.2ghz.
This is the guide I followed: http://www.mediasmarthome.com/article/14459/Upgrading-the-HP-EX485-487-MediaS...
One last thing, if you follow the guide listed above, the author made a mistake when he said to apply the thermal cpu past vertically as pictured. A vertical application is used for Core 2 Duo cpus, but the older e5200 is a Dual Core cpu requiring a horizontal bead of paste.
http://www.arcticsilver.com/intel_application_method.html
EX487 e5200
Anyone try out a Q8300 in their EX485? Micro Center seems to be clearing these out (only $99), and I was going to buy one if anyone else has had luck getting them to work. Micro Center doesn't allow CPU returns, so I don't want to waste $99 if there's little chance of this CPU working.
http://www.microcenter.com/single_product_results.phtml?product_id=0341725
Thanks!
A couple of comments, Jeffla, no doubt there is some confusion about paste application on the E5200 but according to an Intel engineer the design of the E5200 is more like the Xeon 5200 and should have the paste applied in a verticle line. If the paste is properly applied it will migrate and completely cover the top of the CPU.
JTravers, the E5200 at 65w is already pushing the power supply and passive processor cooling to its limits, using a 95w will probably push the platform over the edge leaving you with an intermittent (or worse blown up) unit. (I won't even go into the Quad core issues with the BIOS)
Regards
GJ51 said: 95W Be sure to keep it in the refrigerator.Does the basement count ;)
StarNet said: A couple of comments, Jeffla, no doubt there is some confusion about paste application on the E5200 but according to an Intel engineer the design of the E5200 is more like the Xeon 5200 and should have the paste applied in a verticle line. If the paste is properly applied it will migrate and completely cover the top of the CPU. JTravers, the E5200 at 65w is already pushing the power supply and passive processor cooling to its limits, using a 95w will probably push the platform over the edge leaving you with an intermittent (or worse blown up) unit. (I won't even go into the Quad core issues with the BIOS) Regards
True. It's probably best to wait for one of the 65W quad cores to drop in price. I was just hoping that the 95W rating was conservative and that temps/power draw may not be as bad as suspected since it is a 45nm part.
Thanks!
Does the basement count ;)
Sure does! Especially if you live in Alaska, it's January, and there's a wall missing ... 
I just upgraded my ex485 with a Celeron 3400, and had to set PECI to 01 for it to POST. Now it's going great; except for that shutdown problem, and vmx (Intel virtualization extension) isn't being seen by the OS even thought it is enabled in the bios.
So I'd like to have a play with patching in bits from the ex490 BIOS. However, all of the links to the roms in this thread are dead. Can someone please repost the original BIOS for the ex490 or ex495?
Thanks!
-
For those interested in how the ex475 BIOS was hacked:
http://www.tomshardware.co.uk/mediasmart-ex475-hack,review-31530-9.html
GJ51 said: Just looking on the Intel Processor Spec Finder indicates that the CPUID String for the L3110 and the E5200 and the Q8200S are all 1067Ah. I did a BIOS read on the ex485 and it lists 06F6, 06F2, 06FB, 06FD, and 0661.Bringing back an old but quite informative post from GJ51, I have a E2160 that I've pad modded to run @ 3.0GHz (originally a 1.8GHz part); it is completely stable with no additional cooling other than the stock heatsink/fan (it's currently in a Dell Vostro 200 slim tower). It has a a CPUID string of 06FDh, I believe. Noting what GJ51 wrote above, I wonder if I could put this processor in my EX485 and have it work without any other modifications. In theory, it seems like it should sleep/shutdown fine since the CPUID is already in the EX485's microcode. Anyone see any potential issues with this? If not, I think I might try it out this weekend.
I just tried to upgrade my EX-487 to an E5200 (SLB9T). I flashed the bios with the original chip still installed, rebooted and everything worked fine. I installed the E5200 and got the no hard drive leds with the health led blinking. So I installed the original and plugged in a usb keyboard (to see if it would be recognized) and started the system. Everything was working for about 10 seconds when all the leds when out except the power indicator. Fans are spinning. Holding the power button in has no effect; I have to pull the cord.
Anyone have any clues?
Just upgraded my Ex485 with a E8400. Seems to run good and not any hotter than the stock celeron.
My Slay7 E5200 came in by UPS this afternoon. At 6:30p I did my shutdown and at 8:25p I powered up -- 3 minutes later all the lights were blue!!!
Wow it works! I mean it really really works.
When I remote in my Task Manager shows two CPU's and the System Info shows two CPU.
In the connector it shows 2 cpu's in the system summary and in the Resources Tab it shows "Pentium Dual-Core CPU E5200 @2.5GHz"!!!
I'm still going to get a KVM but the pressure is off. I think it will come in handy down the road.
This thing is really fast.
The only glitch was when I went to pull the heat sink off -- it was stuck and it took forever it seemed to get off -- then I had to clean it off which was an entirely new experience for me but I followed the AS guidelines from their web site including Tinting the heatsink -- seems to be working. My temps peaked at 55 wit both cores going near 100% but now they have settled down to 47-9 degrees -- one or two degrees lower than before with the Celeron 440.
BYW: I got the E5200 Slay7 from Code Micro, Inc.
Pictures of my installation process are at http://homeservershow.com/forums/index.php?/topic/2814-replacing-cpu-on-ex-487/
EX487 4*WD20EARX - 2TB WD Elements for Server Bkup - w/Athena AP-MFATX30 - 4GB-RAM, Pentium Dual-Core CPU E5200 @2.5GHz // Main-Client: Gigabyte GA-X58-UD3R 12GB Ram 1TB SATA6 HD w/Win 7 Pro 64b SP1 & VM WHS-V1
Can someone help me (who has attached KVM to MSS) giving step-by-step instructions how I can change Speed Stepping in BIOS with only USB keyboard connected to MSS? Just like it is described here http://www.mediasmartserver.net/forums/viewtopic.php?t=609 for changing Power Management Settings?
I need it to resolve the same problem after upgrade to E5200 processor:
http://www.mediasmarthome.com/forum/thread/11977/HP-EX-485-CPU-upgrade-sucess...
I'm not a technical guy and I'm quite curious. E5200 is LGA 775 vs. celeron 440 is LGA 1155. So it fits?
The Celeron 440 (Conroe) is a LGA775 chip, not 1155, the only thing that will work in the MSS (480/490 Series) is a 775 based chip. (And not all of those as well)
StarNet said: The Celeron 440 (Conroe) is a LGA775 chip, not 1155, the only thing that will work in the MSS (480/490 Series) is a 775 based chip. (And not all of those as well)Got it & thanks!