I'm no expert but I wonder if you could find a sample of a similar Mac microcode and then compare with the PC version you found and see how they match up or if they're radically different...?
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
iM 4,1-14,4 iMac 21.5" Late 2009 (10,1) - CPU Upgrade E8600
- Thread starter AphoticD
- Start date
- Sort by reaction score
imac 27 late 2009 core 2 duo You can update the graphics card and which one I understand.
thanks
thanks
Maybe a silly question: Can the ones coming with Core2 be upgraded to i5 or i7? Also will this allow 32 GB RAM? (I know the late 2009 with i7 from factory can accept 32 GB (not officially but practically) but the Core2 only 16.)
No and no. i-series CPUs use a different socket and do not physically fit the socket Core 2 CPUs use.
Thank you. What about Xeons? Did somebody tried it in a Dualcore2 iMac?
I know this was done for i-series iMacs with the Xeons corresponding to those.
NO.
Even the Core 2 quad CPUs (LGA 775 or modified LGA 771 sockets) won't be recognized.
What would modifying a core imac into a core-i one require? Is it enough to swap the logicboard?
I replaced the platform B1 microcode with the xeon platform 44 microcode and flashed it. The xeon microcode stops the e7600 from booting, which was to be expected, so I tore it apart and put in the modified x5270 in it (this is tested and working on another 775 board) put it all back together and.................... no boot. Just goes full fan speed on all fans nothing displays, any ideas? Does mac EFI have a FIT table?did you inject the appropriate microcodes into the firmware? without it it might not boot properly espically being an E0 stepping CPU they have some anomalies that need Microcode to fix, infact im suprised it even POSTed without microcode!
you can use @dosdude1's tools to dump the BootROM and inject microcode into it from herehttp://dosdude1.com/apps/
If you got this far, then the problem is not a missing microcode. Updating the microcode is the first thing the processor does during boot, right after reading the CPU ID and before even initializing RAM.
Most likely the problem is somewhere in macOS. You should try booting to Windows, Linux, the EFI shell, the Mac Startup Manager (boot selector), or to maybe some other version of macOS.
Last edited:
Looks like you bricked your iMac. 😥 You may need to use a programming clip to restore the original EFI Rom. But the experiment seems to show that the firmware needs a matching microcode to post.
Both the X5270 and the E8600 share the same CPU ID, 1067A. All the processors made are stepping E0 (SLB**). They should be compatible, except for the microcode.
From the Intel manual:
The X5270 has "Platform Type" 0x44, while the E8600 is 0x11. See
XEON Microcodes added to Socket 775 bios
I'd like to see a collection point (for lack of a better term) for some of the bios that have been moddified for the XEON socket 771 to socket 775 mod. This is made with the understanding that ANY BIOS POSTED IS USED @ YOUR OWN RISK! ANY ATTEMPT TO FLASH A BIOS COULD POSSIBLY END WITH A...
It looks like the platform types are coded as follows:
- 0x01 = 00000001 = LGA775
- 0x04 = 00000100 = LGA771
- 0x08 = 00001000 = PBGA437 for Intel Atom
- 0x10 = 00010000 = LGA775 for 45 nm Penryn
- 0x40 = 01000000 = LGA771 for 45 nm Penryn
The Mac Pro 3.1 has an issue with stepping E0 (SLB**) processors. It only accepts stepping C0 (SLA**). This cannot be the problem here, as both processors are stepping E0.
Heres an update:
I used dosdude's microcode tool to replace the microcode, I have experimented with different changes to the microcode before. Updating, moving or replacing all cause the machine not to boot* (This is with the original cpu). Now heres the catch, it will boot every now and again if you unplug it/leave it for a bit, move the ram to different slots and persist with some varition of those things, it will at some point just randomly boot (chimes and boots into chosen os), I would then go into debian and flash the original efi back. And with the original efi it would be back to working normally as one would expect.
So I decided to try replacing the platform B1 microcode with the same version but with the platform 44 microcode (for xeons), so the location of the microcode in the efi is the same and same version, only now it says platform 44 instead of B1. I flashed it and it didn't boot (no surprise), so I tore it down and put in the x5270. Surely with just the platform change and a cpu and EFI that is the same in every other way it would boot? Well as you know it didnt, so I tried the "unplug it/leave it for a bit, move the ram to different slots and persist with some varition of those things, it will at some point just randomly boot" strategy, and sure enough eventually I got it to chime, but then it would get hung up on the white screen (same as Locla). The second time I got it to chime I was able to get to the boot select but as soon as I selected something it froze. So it obviously wasnt happy.
But then I made some progress thanks to @m0bil , turns out the microcode tool doesnt recalculate the header checksum or the CRC32. So in accordance with @m0bil's testing on the 2011 imac, I changed the crc32 value to 00 00 00 00 to get the bios to skip the CRC check of the microcode volume. Then simply recalculated the header checksum in UEFITool NE, then flashed the new image with a probe.......... and still it wouldnt boot. I think you know whats coming.... more unplugging/leaving it for a bit, moving the ram to different slots until it boots. And eventually I got it to chime and this time I got it to boot into debian WITHOUT IT GETTING HUNG UP! (HDD and screen are not plugged in so fans are loud). Progress??
So then I thought what if I just calculate the crc32 value instead of using the 00 00 00 00 skip thing, so I did the calculation, redid the header checksums, triple checked everything, got the imac to boot into debian again and flashed the image. It restarted fine but got hung up trying to boot debian, so I shut it of and same thing back to the full fan speed. So at this point im not sure what to think.
NOTE: Ive done many PRAM resets on the successful boot, and it will always chime again indicating that the pram has been reset and will continue to boot. Its just once you try restart it again it goes back to the full fan speed and no chime. Perhaps removing the PRAM battery would fix this but it doesnt really seem like a great solution, this shouldnt really be needed.
OTHER LESS IMPORTANT NOTE: When it doesnt chime, it either goes full fan speed no beeping or it gives the ram beep for no ram detected (even though there is RAM in it) not sure what causes it to do one or the other.
So I am at a loss as to why it behaves like this.
I used dosdude's microcode tool to replace the microcode, I have experimented with different changes to the microcode before. Updating, moving or replacing all cause the machine not to boot* (This is with the original cpu). Now heres the catch, it will boot every now and again if you unplug it/leave it for a bit, move the ram to different slots and persist with some varition of those things, it will at some point just randomly boot (chimes and boots into chosen os), I would then go into debian and flash the original efi back. And with the original efi it would be back to working normally as one would expect.
So I decided to try replacing the platform B1 microcode with the same version but with the platform 44 microcode (for xeons), so the location of the microcode in the efi is the same and same version, only now it says platform 44 instead of B1. I flashed it and it didn't boot (no surprise), so I tore it down and put in the x5270. Surely with just the platform change and a cpu and EFI that is the same in every other way it would boot? Well as you know it didnt, so I tried the "unplug it/leave it for a bit, move the ram to different slots and persist with some varition of those things, it will at some point just randomly boot" strategy, and sure enough eventually I got it to chime, but then it would get hung up on the white screen (same as Locla). The second time I got it to chime I was able to get to the boot select but as soon as I selected something it froze. So it obviously wasnt happy.
But then I made some progress thanks to @m0bil , turns out the microcode tool doesnt recalculate the header checksum or the CRC32. So in accordance with @m0bil's testing on the 2011 imac, I changed the crc32 value to 00 00 00 00 to get the bios to skip the CRC check of the microcode volume. Then simply recalculated the header checksum in UEFITool NE, then flashed the new image with a probe.......... and still it wouldnt boot. I think you know whats coming.... more unplugging/leaving it for a bit, moving the ram to different slots until it boots. And eventually I got it to chime and this time I got it to boot into debian WITHOUT IT GETTING HUNG UP! (HDD and screen are not plugged in so fans are loud). Progress??
So then I thought what if I just calculate the crc32 value instead of using the 00 00 00 00 skip thing, so I did the calculation, redid the header checksums, triple checked everything, got the imac to boot into debian again and flashed the image. It restarted fine but got hung up trying to boot debian, so I shut it of and same thing back to the full fan speed. So at this point im not sure what to think.
NOTE: Ive done many PRAM resets on the successful boot, and it will always chime again indicating that the pram has been reset and will continue to boot. Its just once you try restart it again it goes back to the full fan speed and no chime. Perhaps removing the PRAM battery would fix this but it doesnt really seem like a great solution, this shouldnt really be needed.
OTHER LESS IMPORTANT NOTE: When it doesnt chime, it either goes full fan speed no beeping or it gives the ram beep for no ram detected (even though there is RAM in it) not sure what causes it to do one or the other.
So I am at a loss as to why it behaves like this.
Last edited:
You were both using a Xeon X5270 processor, with a max TDP of 80 watts. The crashes may simply happen because the processor is drawing too much power from the logic board. The supported CPUs have a max TDP of 65 watts.
One should retry this with lower powered Xeons, like the E5240 (65W), L3110 (45W), L5240 (40W), or the L5430 (50W).
Products formerly Wolfdale
Products formerly Wolfdale product listing with links to detailed product features and specifications.
ark.intel.com
As per my question elsewhere, can anyone tell me if these mods, particularly the E8600, actually allow the RAM to be run at 1333MHz?
I know im a couple years late but you mentioning it beeping at you for RAM not detect is interesting
what RAM do you have in the system? MCP79 Macs are quite picky about which RAM they will work with (and on top of that they have a bug, where they will read the highest speed from the SPD data and try to run at that, so you install 1333Mhz RAM, and rather then the system just running that ram at the 1066Mhz the system is rated for, it actually tries to overclock itself to run at 1333Mhz, and most of the time it will fall over doing so)
so just as a sanity check was the system fully stable even with the E8600 in it? i'd make sure your using PC8500 RAM for testing, just 1 or 2 small sticks to remove any and all RAM related funky-ness
just thought i'd mention it just incase....
I was using a single 1066 dimm for testing. I actually had some success booting into windows though after using both 771 and 775 microcode's, and posted this: https://forums.macrumors.com/thread...-duo-cpu-upgrades-x5270-q9550s-e5450.2371669/
About a week after that post, I gave up and reassembled the computer for a final time with the e8600 and it has resided in my workshop ever since.
Annoyingly though I had a lightbulb moment literally a day later on how to make it work. But I still haven't got around to testing it as I can't be ****ed opening it up again as it works nicely as a workshop computer.
But if someone else wants to try it, I can modify their efi and they can test it (I take no responsibility for bricked imac).
just found this thread, very cool getting to boot into windows
i actually have an e5450 in a dell with the pin mod and was thinking of swapping it into the mac
but its a lot of work to open it up.
I am currently running windows 10 with 4gb of ram lol but its not bad.
The mxm gpu is the ati 4670 (do you know which other card i can upgarde off hand? i know the vbios mods dont work on this model)
nice work on the mods though
The 4670 is the only reliable option that I'm aware of. And good luck with the e5450 some bios mods are almost certainly going to be needed.