Same S-spec. There is only one non-engineering sample S-spec for X5680.
For Nehalem and Westmere Xeons, matched pair is a sales tactic.
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.
MP All Models Mac Pro CPU Compatibility List
- Thread starter ActionableMango
- WikiPost WikiPost
- Start date
- Sort by reaction score
- Status
For Nehalem and Westmere Xeons, matched pair is a sales tactic.
I have a MacPro1,1 with dual 5150 flashed to MacPro2,1. I've already also done the SMC flash too.
I was thinking of getting the 80 W G0 stepping E5345 SLAEJ but it looks like I'll have to buy it from overseas. I have no idea how long it will take to get here. However, it turns out there is a seller nearby in my country with the 120 W G0 stepping X5355 SLAEG available. I could probably get it within a week since there is no international shipping and customs to worry about. Pricing is similar.
Just how much better is the power utilization of X5355 G0 SLAEG compared to the other X5355 steppings in real world usage? I read some complaints about heat for the X5355 B3, but that one idles higher.
Most of the usage will be relatively light (surfing, etc.) but with some video decoding (720p Netflix macOS or else 1080p Netflix Windows, and 1440p YouTube VP9). No video editing or stuff like that. BTW, I can already do this stuff with the 5150, but there are occasional stutters and pauses when multi-tasking that I'd like to get rid of. So, I'm thinking I'd probably be OK 95% of the time in terms of fan noise with the X5355. However, probably based on paper specs, E5345 would be the ideal solution. I even contemplated L5335, but that's more expensive, and I'd have to get it overseas too.
EDIT:
Too late. I've already ordered the E5345 SLAEJ pair from overseas. US$20 shipped. Besides the previous complaints by some of fan speed/noise with some X5355 upgrades, it was also because of some comparisons I came across.
www.ehmac.ca
Note though that this X5355 is the B3 SLAC4, not the G0 stepping.
Also, I came across a Dell Server comparison, comparing their own E5345 server vs. their own X5355 server.
I don't know what steppings these are, but nonetheless the comparisons gave me pause on the X5355. I guess this shouldn't come as a surprise since the E5345 has an 80 W TDP and the X5355 has a 120 W TDP. Sure the G0 stepping of the X5355 idles closer to 25 W, but I was hoping for better behaviour at high load. Most of the time my usage won't stress the machine, and both chips would handle that fine. However, when the load goes up, it'd be nice if the fans could stay quiet longer.
Does anyone know the idle power consumption of E5345? FWIW, this review has the E5345 idling at the same Wattage as 5160, but the table in the first post of this thread doesn't listing the 5160 idle power either.
www.anandtech.com
Interestingly though, the 5160 at load uses more power than E5345, despite the fact both are 80 W TDP chips, and 5160 only has two cores.
I was thinking of getting the 80 W G0 stepping E5345 SLAEJ but it looks like I'll have to buy it from overseas. I have no idea how long it will take to get here. However, it turns out there is a seller nearby in my country with the 120 W G0 stepping X5355 SLAEG available. I could probably get it within a week since there is no international shipping and customs to worry about. Pricing is similar.
Just how much better is the power utilization of X5355 G0 SLAEG compared to the other X5355 steppings in real world usage? I read some complaints about heat for the X5355 B3, but that one idles higher.
Most of the usage will be relatively light (surfing, etc.) but with some video decoding (720p Netflix macOS or else 1080p Netflix Windows, and 1440p YouTube VP9). No video editing or stuff like that. BTW, I can already do this stuff with the 5150, but there are occasional stutters and pauses when multi-tasking that I'd like to get rid of. So, I'm thinking I'd probably be OK 95% of the time in terms of fan noise with the X5355. However, probably based on paper specs, E5345 would be the ideal solution. I even contemplated L5335, but that's more expensive, and I'd have to get it overseas too.
EDIT:
Too late. I've already ordered the E5345 SLAEJ pair from overseas. US$20 shipped. Besides the previous complaints by some of fan speed/noise with some X5355 upgrades, it was also because of some comparisons I came across.
Upgrading the CPU of a Mac Pro 1.1
E5345 owner said:
X5355 owner said:
Note though that this X5355 is the B3 SLAC4, not the G0 stepping.
Also, I came across a Dell Server comparison, comparing their own E5345 server vs. their own X5355 server.
I don't know what steppings these are, but nonetheless the comparisons gave me pause on the X5355. I guess this shouldn't come as a surprise since the E5345 has an 80 W TDP and the X5355 has a 120 W TDP. Sure the G0 stepping of the X5355 idles closer to 25 W, but I was hoping for better behaviour at high load. Most of the time my usage won't stress the machine, and both chips would handle that fine. However, when the load goes up, it'd be nice if the fans could stay quiet longer.
Does anyone know the idle power consumption of E5345? FWIW, this review has the E5345 idling at the same Wattage as 5160, but the table in the first post of this thread doesn't listing the 5160 idle power either.
AMD's Opteron hits 3.2GHz
Interestingly though, the 5160 at load uses more power than E5345, despite the fact both are 80 W TDP chips, and 5160 only has two cores.
Last edited:
Hello to all, I have a Mac Pro Early 2009, with dual cpu 2.66MHz. Some year ago, I upgraded firmware following guides here on macrumors to update the Mac Pro 2010 (5.1).
That's what I did.
On December, I decided to upgrade several components:
Dual CPU X5690 delidded, 3.46GHz, bought here.
Sapphire Pulse RadeOn RX 580 8GB
Everything worked: Mojave recognizes everything.
Previously, during some update, I lost my recovery partition, so it's impossible to run Diagnostics, safe mode, recovery and so on. Plus, RX 580 didn't show up boot screens, even if I succesfully upgraded firmware from 138.0.0.0 to 144.0.0.0.
I didn't upgrade RAM: I left inside 1066MHz modules.
After installing the new CPUs, suddenly the mac pro restarts randomly after some minute. But restarting, the mac fully worked without other reboots.
This scenario was valid until 15 days ago. After that, Mojave suddenly restarts, but when booting the OS doesn't start, or if it starts, the system freezes and reboots again, continuously.
It's impossible to have a stable machine.
I tried everything: restoring the old GPU unit (ATI Radeon 5770), removing ram modules from 40GB to 16GB, changing HD with a fresh installation disk: after some minute, the mac freezes, sometimes restarts, sometimes remains freezed.
Today I tried the last thing: removed the X5690 cpus and inserted the original chips.
Mojave starts, it was on since 3 hours, without freezes or reboots. But right now I heared a "stuck" sound and the system restarted.
I suspect many things, but I don't know how to solve:
the cpus X5690 are defective (15 days ago the reseller sent to me another CPU in warranty, but doesn't change the result). But restarts occurs even with old CPUs.
RAM are 1066MHz and the new CPU cannot work with these types.
Mac pro 4,1 migrated to 5,1 doesn't have sufficient ventilation and the CPUs stucks after some minute, causing reboots.
Mojave installation (upgraded from High Sierra) is corrupted (though disk utility says there is no problem on the disk).
Logic Board or/and processor tray is/are gone.
Any type of help is very appreciated.
That's what I did.
On December, I decided to upgrade several components:
Dual CPU X5690 delidded, 3.46GHz, bought here.
Sapphire Pulse RadeOn RX 580 8GB
Everything worked: Mojave recognizes everything.
Previously, during some update, I lost my recovery partition, so it's impossible to run Diagnostics, safe mode, recovery and so on. Plus, RX 580 didn't show up boot screens, even if I succesfully upgraded firmware from 138.0.0.0 to 144.0.0.0.
I didn't upgrade RAM: I left inside 1066MHz modules.
After installing the new CPUs, suddenly the mac pro restarts randomly after some minute. But restarting, the mac fully worked without other reboots.
This scenario was valid until 15 days ago. After that, Mojave suddenly restarts, but when booting the OS doesn't start, or if it starts, the system freezes and reboots again, continuously.
It's impossible to have a stable machine.
I tried everything: restoring the old GPU unit (ATI Radeon 5770), removing ram modules from 40GB to 16GB, changing HD with a fresh installation disk: after some minute, the mac freezes, sometimes restarts, sometimes remains freezed.
Today I tried the last thing: removed the X5690 cpus and inserted the original chips.
Mojave starts, it was on since 3 hours, without freezes or reboots. But right now I heared a "stuck" sound and the system restarted.
I suspect many things, but I don't know how to solve:
the cpus X5690 are defective (15 days ago the reseller sent to me another CPU in warranty, but doesn't change the result). But restarts occurs even with old CPUs.
RAM are 1066MHz and the new CPU cannot work with these types.
Mac pro 4,1 migrated to 5,1 doesn't have sufficient ventilation and the CPUs stucks after some minute, causing reboots.
Mojave installation (upgraded from High Sierra) is corrupted (though disk utility says there is no problem on the disk).
Logic Board or/and processor tray is/are gone.
Any type of help is very appreciated.
Last edited:
-CPUs: I think you answered the question already: same issues with other CPUs > not the CPU
-RAM will work fine with Westmere (but at 1066 MHz). If you'd buy 1333MHz RAM, it would work at 1333MHz with the 6core CPU and at 1066 MHz with the OEM CPUs.
-Ventilation is identical for 4,1 and 5,1
-OS corrupted: just read you already tested a fresh installed OS so: not the issue.
-LogicBoard defective: could be, but there are lots of possiblities (e.g.PSU)
-Did you install e.g.MacsFanControl or similar to setup the fans and control the Temps?(if possible install it and post a screenshot of the Temps)
-Did you already check the NorthBridge heatsink pivets? (they can break after so many years and cause overheating the NorthBridge)
-Regular cleaning and de-dusting the MP completely (in combination with a fan control app) is highly recommended to keep the MP 4,1 /5,1 alive and avoid overheating parts especially the NB.
- the cause could be a lot of things ,I'd start with installing MacsFanControl to see the Temps and visually check the NB pivets (they are partly underneath heatsink CPU A).
-the HD 5770 is not compatible with Mojave, so what OS did you install/run?
In case of further troubleshooting etc its recommended to download the Technician Guide :
BTW: by default the RX 580 doesnt show the bootscreen , to get a bootscreen you could:
-have it flashed /buy a flashed RX580
-or install Open Core
Last edited:
First of all thank you for your quick answer.
Yesterday I installed MacsFanControl. Before reboot, I notice that CPU A went to 75 Celsius degrees, while CPU B around 45/50 Celsius degrees. So I unmounted everything, cleaned the heat sinks, reapplied the thermal paste to seethe differences.
After that, I noticed that CPU A went to 50/55 Celsius degrees, while CPU B 45/50.
Though while unmounting the heat sink of CPU A, I noticed that the northbridge pivets is fixed with only one screw. The other screw was broken on December while unmounting the heat sink. But I tried this morning to attach the heat sink while maintaining the northbridge pivets alligned.
For HD 5770 I really don’t know why is working! I bought RX 580 to install Mojave (upgraded from High Sierra). Now that I tried to boot Mojave with the old graphics card, maybe it can magically starts because there are drivers installed previously from High Sierra? Maybe Mojave controls the correct GPU while installing and not while it has been correctly installed?
I think I’ll install OpenCore to see boot screens, but after resolving the reboots problem.
Maybe installing this firmware it suddenly restarts and I loose the Mac forever.
So it can be the lost screw for the northbridge pivets? After installing the heat sink on the cpu I can see that the northbridge pivets is stable and not moving.
It's essential that the NB heatsink is properly connected (with thermal paste) to the NB die.
The NB can cook easily without proper cooling...
What are the Temps for NB diode and NB heatsink?
Delta NB diode and NB heatsink should be ideally under 10 degrees, if more than ~12 degrees I'd repaste and remount the NB heatsink .
If NB diode is above ~75 Celsius shut the MP down and dont use it until you got the NB heatsink properly placed.
There are lots of opinions about the pivets , personally I install them with the original springs(important!) and (nylon) M3x16mm bolts and nuts.
There's a kit on eBay available too:
Mac Pro 4,1 and 5,1 Northbridge heatsink retainer spring kit | eBay
This precision spring kit is the definitive answer to repairing your Mac Pro 4,1 and 5,1 Northbridge heat sink broken rivets. Fix your Mac Pro before you have a catastrophic failure like many have already.
www.ebay.com
After all it's a miracle your MP survived for so long without the heatsink correctly seated...
To further troubleshoot I'd install e.g. High Sierra because of the 5770.
Although it booted afaik this is failmode.
BTW: I just read it's "rivets" not pivets , sorry
Thank you very much for your exhaust post! I just ordered the clips on eBay, though not that metal based, but plastic. Because I live in Italy and those you advise are in Canada (means that arrive here at the end of august or September, too much time).
The clips I ordered are these.
Hope that I didn’t fry the NB! Now the MP is turned off until the clips arrives (Should arrive in a week). Then I’ll insert the X5690 again, repaste everything and I cross my fingers.
RadeOn 5770 I used for boot screen only, I’ll install the RX 580. If everything goes well, I think I’ll install OpenCore and Mojave on top of the existing to regenerate the recovery partition.
I’ll post again when everything is done.
Thank God MP2019 arrived last week, and there are a MacMini and a MBP. Work is not compromised ?
The clips I ordered are these.
Hope that I didn’t fry the NB! Now the MP is turned off until the clips arrives (Should arrive in a week). Then I’ll insert the X5690 again, repaste everything and I cross my fingers.
RadeOn 5770 I used for boot screen only, I’ll install the RX 580. If everything goes well, I think I’ll install OpenCore and Mojave on top of the existing to regenerate the recovery partition.
I’ll post again when everything is done.
Thank God MP2019 arrived last week, and there are a MacMini and a MBP. Work is not compromised ?
No problem, glad to help you!
Unfortunately the clips you ordered are not correct...
If you still have the springs, you can just use M3X16mm nylon or metal screws and bolts.
The springs are essential to maintain the (correct) pressure on the die.
Otherwise I'd wait for the screws from Canada, or remove the springs from the clips you ordered and buy some M3X 16mm screws and nuts...
I recommend to make a TM backup of the existing disk and (after installing OC) make a fresh install, after that restore with immigration assistant.
In Bocca Lupo....
Unfortunately the clips you ordered are not correct...
If you still have the springs, you can just use M3X16mm nylon or metal screws and bolts.
The springs are essential to maintain the (correct) pressure on the die.
Otherwise I'd wait for the screws from Canada, or remove the springs from the clips you ordered and buy some M3X 16mm screws and nuts...
I recommend to make a TM backup of the existing disk and (after installing OC) make a fresh install, after that restore with immigration assistant.
In Bocca Lupo....
The reseller told me that other people with Macs used without problems
I have 1 spring only, currently mounted to fix the heatsink on the left. The other was broken on December and I thrown it away.
Really? Before buying, I read this post, and I followed up the ebay link. That item is no longer available, but I can see specifications on the pictures. That ones I ordered are identical in dimensions and form factor. That's why I bought them.
I have a TM backup on that MP, right now I'm cloning the SSD with mojave of that MP on my MP2019 to a new drive, to be sure. And you convinced me, I'll do a fresh Mojave install with a migration assist. I need to buy another SSD to make this operation. Those others I have doesn't have sufficient space to host the original MP's Mojave (700GB on a 1TB SSD).
Crepi!In Bocca Lupo....
From what I've seen in the description of the clips you ordered, I assume they will not fit, because the heads are too high and one of them may touch the CPU heatsink or even worse avoid the correct position of the CPU heatsink...
Therefore I'd recommend to remove the springs from the clips you ordered and buy some M3X16mm screws ,nuts and washers (nylon or metal)....
To get an impression of the correct seating I attached some pictures.
In one of the pictures you can see that the heads of the original clips are very close to the heatsink (note the dust: this is a MP 4,1 I recently bought and upgraded).
Therefore I'd recommend to remove the springs from the clips you ordered and buy some M3X16mm screws ,nuts and washers (nylon or metal)....
To get an impression of the correct seating I attached some pictures.
In one of the pictures you can see that the heads of the original clips are very close to the heatsink (note the dust: this is a MP 4,1 I recently bought and upgraded).
Attachments
So, do you installed regular M3x16mm screws, nuts and washers with the original springs? Because in this case I can buy them in a regular and strong furnished hardware store.
Yes, I've seen lots of people doing it with ordinary metal screws.
The eBay kit is metal too...
The reason I choose nylon is that the first time I did it, I wasnt sure if there would be any electrical path on the underside of the pcb close to the nut of the screw.
Oh sh*t! 15 (!!!) bucks for two 3mm screws with self locking nuts and springs.
Should be 1.5 instead!
I would like to thank again KeesMacPro for helping me to find the problem of my MP. 10 days searching around, unmounting every single component of the mac, trying to remove RAM, remove CPUs, remove HDs...
It was Northbridge Die that has a broken rivet on the right side. It's a miracle that it's still alive: that little piece was broken on december, and the Mac worked a lot from that date, with some reboots.
The thermal paste on the NB die was so encrusted: it took me 10 minutes with the arctic cleaner to remove it.
I repasted cpus and the NB.
Now it's working at 100%, it's switched on since 12 hours, without any reboot.
I discovered that some RAM Module was damaged too (red lights on cpu tray), but once removed, the red lights switched off.
I bought two M3x30 screws and nuts, because I found (!) the little spring of the broken screw on december.
I sawed the long screws. Now the heatsink is stable on the NB die. They are very similar to those here in attachment, but I bought them in a regular hardware store (20 pieces, 1 💶).
I hope that this post can help other people having the same problem.
Attachments
Last edited:
Yes, is likely the best combination, using the stock-springs as they are stronger than most of the other rivets's. I'm afraid, including the 15-bucks-ebay-set's.
It's exactly what i use. Only difference: As i was not able to find selflocking 3mm nuts at my location, i use two normal nuts locked against each other on each screw. Total costs of somewhat around 4 euros. But, at least for the screws and nuts, i have 8 "sets" left in spare. Might eventually sell them on ebay. 10 for each should be ok without the springs... 😆
if someone else is looking for new northbridge clips those work great have the exact original dimensions and their springs are strong enough to get perfect cooling numbers (delta tdiode/heatsink).
cheers
tom
Delock 18273 Push Rivet Niete
Kurzbeschreibung Dieses Schnappnieten Set von Delock dient zur Befestigung von einem Kühlkörper auf der Platine.Spezifikation Schnappniete für Kühlkörper Material: Kunststoff Feuerwiderstandsklass...
www.amazon.de
cheers
tom
Attachments
Last edited:
we have 6 mac pro 4.1/8/2.26 16gb hd5770 and upgraded them to 5.1/12/3.06 96gb wx7100
Attachments
Last edited:
Another Mac Pro 2006 upgrade success today! Thanks for the CPU list guys.
I also had the chance to test this E5345-upgraded MacPro1,1 next to a stock MacPro2,1 with Xeon X5365 (120 W). The latter machine was pulled out of the garbage. I also compared against the baseline Xeon 5150 (65 W).
Note: In order to use the iFixit toolkit to remove the heatsinks, I had to trim off the widest part of the collar from the screwdriver extension. Otherwise it wouldn't fit.
Specs:
MacPro1,1 flashed to 2,1
Boot ROM MP21.007F.B06
SMC 1.15f3
2 x 2.66 GHz dual-core Xeon 5150 (SL9RU - B2 stepping)
8 GB RAM - Kingston 2 x 4 GB server RAM (without the proper heatsinks)
Broadcom BCM94321MC WiFi
No Bluetooth
Apple USB keyboard and mouse
Radeon HD 5770 Mac
MacPro1,1 flashed to 2,1
Boot ROM MP21.007F.B06
SMC 1.15f3
2 x 2.33 GHz quad-core Xeon E5345 (SLAEJ - G0 stepping)
8 GB RAM - Kingston 2 x 4 GB server RAM (without the proper heatsinks)
Broadcom BCM94321MC WiFi
No Bluetooth
Apple USB keyboard and mouse
Radeon HD 5770 Mac
MacPro2,1
Boot ROM MP21.007F.B08 <-- What's the difference with this firmware?
SMC 1.15f3
2 x 3.0 GHz quad-core Xeon X5365 (SLAED* - G0 stepping)
12 GB RAM - OWC 2 x 4 GB, Nuimpact 2 x 2 GB (all with Apple-approved heatsinks)
Broadcom BCM94322MC WiFi
Cambridge Silicon Radio Bluetooth 2.0 module
Apple Bluetooth Magic Keyboard and Magic Mouse
Radeon HD 5770 Mac
*I presume it's SLAED. I didn't disassemble the machine to check, but CPU-Z shows it is G0 stepping.
All had the same original Apple 160 GB HD (Lion) and Crucial BX500 120 GB SSD (El Capitan).
BTW, using a 1,1 HD sled in true 2,1 made for lots of vibration noise. The sled screws are different.
Results:
For Chrome 84 viewing of Costa Rica 4K, I could not smoothly play back 1440p60 VP9 8-bit on any of the setups. It got closer on the 3.0 GHz 8-core, but it still stuttered frequently. On the 2.33 GHz 8-core it stuttered more. Interestingly though, both of them could play 1440p60 fine in Firefox Legacy 67, albeit with a little more headroom on the 3.0 GHz. I'm not sure why Chrome and Firefox differ so greatly here. (As you can imagine, 1440p60 was a total lost cause on the 2.66 GHz 4-core in any browser.)
Cinebench R20 was pretty much as expected:
1248: 3.0 GHz 8-core X5365
980: 2.33 GHz 8-core E5345
546: 2.66 GHz 4-core 5150
The most enlightening finding for me though was the power usage. The results were not unexpected, but nonetheless, the increase in power utilization was huge for X5365.
2.66 GHz 4-core - 2 x Xeon 5150 (TDP 65 W)
6 W: Sleep
178 W: Idle, with Lion HD powered up
171 W: Idle, with Lion HD powered down
272 W: Peak draw during boot
264 W: Peak draw during Cinebench R20
2.33 GHz 8-core - 2 x Xeon E5345 (TDP 80 W)
6 W: Sleep
169 W: Idle, with Lion HD powered up
287 W: Peak draw during boot
287 W: Peak draw during Cinebench R20
3.0 GHz 8-core - 2 x Xeon X5365 (TDP 120 W)
6 W: Sleep
202 W: Idle, with Lion HD powered up
371 W: Peak draw during boot
391 W: Peak draw during Cinebench R20
So, despite the higher TDP, the 2 x 2.33 GHz E5345 actually idled cooler than the stock 2 x 2.66 GHz 5150. However, the idle power of the 2 x 3.0 GHz X5365 was significantly higher, at 33 Watts more than 2 x E5345. (Note that the X5365 machine had a Bluetooth module and 4 GB more RAM.)
Even more striking was the load power. E5345 was marginally higher than 5150, but X5365 was in a totally different league. Its peak power was over 100 W more (!) than either E5345 or 5150. Perhaps a bit of that was due to the higher amount of memory in that machine, but likely the vast majority of the difference in peak power utilization comes down to the CPUs.
One last thing: I had read somewhere that in El Capitan, the flashed 1,1 can be unstable with 512 MB or 1 GB DIMMs installed. I didn't quite believe it, but it locked up once with those DIMMs in El Capitan when I was doing the testing above. Once I removed those DIMMs, all was perfect.
I also had the chance to test this E5345-upgraded MacPro1,1 next to a stock MacPro2,1 with Xeon X5365 (120 W). The latter machine was pulled out of the garbage.
I also compared against the baseline Xeon 5150 (65 W).Note: In order to use the iFixit toolkit to remove the heatsinks, I had to trim off the widest part of the collar from the screwdriver extension. Otherwise it wouldn't fit.
Specs:
MacPro1,1 flashed to 2,1
Boot ROM MP21.007F.B06
SMC 1.15f3
2 x 2.66 GHz dual-core Xeon 5150 (SL9RU - B2 stepping)
8 GB RAM - Kingston 2 x 4 GB server RAM (without the proper heatsinks)
Broadcom BCM94321MC WiFi
No Bluetooth
Apple USB keyboard and mouse
Radeon HD 5770 Mac
MacPro1,1 flashed to 2,1
Boot ROM MP21.007F.B06
SMC 1.15f3
2 x 2.33 GHz quad-core Xeon E5345 (SLAEJ - G0 stepping)
8 GB RAM - Kingston 2 x 4 GB server RAM (without the proper heatsinks)
Broadcom BCM94321MC WiFi
No Bluetooth
Apple USB keyboard and mouse
Radeon HD 5770 Mac
MacPro2,1
Boot ROM MP21.007F.B08 <-- What's the difference with this firmware?
SMC 1.15f3
2 x 3.0 GHz quad-core Xeon X5365 (SLAED* - G0 stepping)
12 GB RAM - OWC 2 x 4 GB, Nuimpact 2 x 2 GB (all with Apple-approved heatsinks)
Broadcom BCM94322MC WiFi
Cambridge Silicon Radio Bluetooth 2.0 module
Apple Bluetooth Magic Keyboard and Magic Mouse
Radeon HD 5770 Mac
*I presume it's SLAED. I didn't disassemble the machine to check, but CPU-Z shows it is G0 stepping.
All had the same original Apple 160 GB HD (Lion) and Crucial BX500 120 GB SSD (El Capitan).
BTW, using a 1,1 HD sled in true 2,1 made for lots of vibration noise. The sled screws are different.
Results:
For Chrome 84 viewing of Costa Rica 4K, I could not smoothly play back 1440p60 VP9 8-bit on any of the setups. It got closer on the 3.0 GHz 8-core, but it still stuttered frequently. On the 2.33 GHz 8-core it stuttered more. Interestingly though, both of them could play 1440p60 fine in Firefox Legacy 67, albeit with a little more headroom on the 3.0 GHz. I'm not sure why Chrome and Firefox differ so greatly here. (As you can imagine, 1440p60 was a total lost cause on the 2.66 GHz 4-core in any browser.)
Cinebench R20 was pretty much as expected:
1248: 3.0 GHz 8-core X5365
980: 2.33 GHz 8-core E5345
546: 2.66 GHz 4-core 5150
The most enlightening finding for me though was the power usage. The results were not unexpected, but nonetheless, the increase in power utilization was huge for X5365.
2.66 GHz 4-core - 2 x Xeon 5150 (TDP 65 W)
6 W: Sleep
178 W: Idle, with Lion HD powered up
171 W: Idle, with Lion HD powered down
272 W: Peak draw during boot
264 W: Peak draw during Cinebench R20
2.33 GHz 8-core - 2 x Xeon E5345 (TDP 80 W)
6 W: Sleep
169 W: Idle, with Lion HD powered up
287 W: Peak draw during boot
287 W: Peak draw during Cinebench R20
3.0 GHz 8-core - 2 x Xeon X5365 (TDP 120 W)
6 W: Sleep
202 W: Idle, with Lion HD powered up
371 W: Peak draw during boot
391 W: Peak draw during Cinebench R20
So, despite the higher TDP, the 2 x 2.33 GHz E5345 actually idled cooler than the stock 2 x 2.66 GHz 5150. However, the idle power of the 2 x 3.0 GHz X5365 was significantly higher, at 33 Watts more than 2 x E5345. (Note that the X5365 machine had a Bluetooth module and 4 GB more RAM.)
Even more striking was the load power. E5345 was marginally higher than 5150, but X5365 was in a totally different league. Its peak power was over 100 W more (!) than either E5345 or 5150. Perhaps a bit of that was due to the higher amount of memory in that machine, but likely the vast majority of the difference in peak power utilization comes down to the CPUs.
One last thing: I had read somewhere that in El Capitan, the flashed 1,1 can be unstable with 512 MB or 1 GB DIMMs installed. I didn't quite believe it, but it locked up once with those DIMMs in El Capitan when I was doing the testing above. Once I removed those DIMMs, all was perfect.
Last edited:
in order to use the cpu to its maximum capacity it is easier and better to use this code in terminal. one 'yes' for every hypertreading core - eg for an real 8-core:
Code:
yes > /dev/null & yes > /dev/null & yes > /dev/null & yes > /dev/null & yes > /dev/null & yes > /dev/null & yes > /dev/null & yes > /dev/null & yes > /dev/null & yes > /dev/null & yes > /dev/null & yes > /dev/null & yes > /dev/null & yes > /dev/null & yes > /dev/null & yes > /dev/null
thats interesting, because the latest publicly available MP2,1 firmware I have come across, does not actually have microcode for the G0 CPU's (they obviously works fine, but does not have microcodes for it)
and all factory MP2,1's I have come across, have B3 stepping CPU's
so to see one with G0 and later firmware then publicly available is very interesting
if you could dump the bootrom and check what microcodes it contains with @dosdude1's tools that would be appreciated
what is the firmware date reported as in CPU Z on the motherboard tab?
Yes that’s true but I wasn’t really wanting to do a temp stress test. Perhaps I’m mistaken, but I’m thinking something like Cinebench or Handbrake may be more representative of real world CPU power draws.
OK I’ll try to do that later today or tomorrow.thats interesting, because the latest publicly available MP2,1 firmware I have come across, does not actually have microcode for the G0 CPU's (they obviously works fine, but does not have microcodes for it)
and all factory MP2,1's I have come across, have B3 stepping CPU's
so to see one with G0 and later firmware then publicly available is very interesting
if you could dump the bootrom and check what microcodes it contains with @dosdude1's tools that would be appreciated
what is the firmware date reported as in CPU Z on the motherboard tab?
Everymac states that the introduction date of this model was 2007-04-04. I'm not sure how to read the CPU-Z date but I'm assuming it's the Eurasian standard, so mainboard manufacture date is 2007-06-09. That means this particular unit likely wouldn't have been available at retail until summer or fall 2007.
ROM dump to follow.
Personally I always write my numerical format dates as YYYY-MM-DD so everyone in the world can understand.
MP21.007F.B08 firmware microcodes:
Is it correct to assume that last 6FB entry isn't there in the B06 firmware?
BTW, do you usually run these machines with SIP enabled or disabled? I only realized my install of El Capitan had SIP enabled when ROMTool told me to disable it.
Is it correct to assume that last 6FB entry isn't there in the B06 firmware?
BTW, do you usually run these machines with SIP enabled or disabled? I only realized my install of El Capitan had SIP enabled when ROMTool told me to disable it.
Last edited: