I'd like to see a credit for MacEFIRom and Netkas, as without the linked firmware updater, much of what is possible would not be possible.
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
Good to know. All I can do is report what I've got. Hope this info helps someone!
Thank you. That's an embarrassing omission on my part, and I've fixed it.
A note - the Clovertown CPUs in the chart are all 4 Core CPUs, the L5335s, L5320s, and E5320s are listed as "2 Core" currently.
Is there any noticeable difference between the stock 2.8 E5462 in the 3,1 Mac Pro and the 3.16 X5460?
Hard to say, same can be questioned for every CPU that is close to another in terms of cores / clockspeed. I don't think it would be noticeable in everyday use but it would shave off some seconds for other full load tasks.
Thanks - I don't think I'll end up bothering as the costs (and risks of breaking something) is greater than the marginal benefit of a slightly better CPU. I was just curious!
Depends on usage. In single threaded tasks it's 10% faster than 5462, in multi threaded about 8%.
There's no practical sense in upgrading 3,1 CPUs at all, except the situation when you have an uniprocessor one.
Here gain is pretty good.
https://browser.primatelabs.com/geekbench3/1835159
The good link's CPU sell fast. Anyway, some users suggest that get a whole used server and only extract the CPU is cheaper than just buy the CPU. And the CPU in the server most likely never be OCed, always have proper cooling, etc. You can also sell the parts you don't need, which get some money back.
Good point. But this means also researching what servers contain what CPU. Maybe worth investigating, but as a Mac user since too many years my knowledge of other servers is very poor
In single-threaded performance the fastest 3,1 processor the 3.2 GHz X5482 is 98% of the performance of the fastest Apple configured 5,1 (3.07GHz X5675) & 90% of the performance of fastest possible 5,1 CPU (3.46GHz X5690).
This demonstrates two things, firstly that the 2008 Mac Pro 3,1 still has very decent performance for today's applications & secondly that Moore's Law ran out of steam some years ago in terms of single stream performance.
Unfortunately, you can't compare GHz for GHz in that manner. There are lots of other things that figure in to the equation. Here's 3GHz vs 2.93GHz...
Agreed. I owned a 3,1 and then a 4,1->5,1 and there are pronounced differences in logic board, memory speed, PCIE lanes, etc. that all impact performance. Even the CPUs are significantly different in that they support virtual cores.
I wasn't comparing GHz as it doesn't make sense with different architectures. I was quoting benchmark figures for the Xeon processors.
CPUMark
Intel Xeon X5482 @ 3.20GHz Single Thread Rating: 1379
http://www.cpubenchmark.net/cpu.php?cpu=Intel+Xeon+X5482+@+3.20GHz
Intel Xeon X5675 @ 3.07GHz Single Thread Rating: 1405
http://www.cpubenchmark.net/cpu.php?cpu=Intel+Xeon+X5675+@+3.07GHz
Intel Xeon X5690 @ 3.47GHz Single Thread Rating: 1520
http://www.cpubenchmark.net/cpu.php?cpu=Intel+Xeon+X5690+@+3.47GHz
Incidentally those two Geekbench results are not comparable as they are two different versions of Geekbench & two different versions of OS X. I was also quoting the X5482 2.2GHz part which has a Geekbench single core score of 1872 which is still 87% of the single core performance of the 2.93GHz X5570 you quote.
Bottom line the single threaded application performance of the Mac Pro 3,1 is still pretty decent compared to the 5,1 & single threaded CPU performance barely increased in five years.
The 3,1 CPUs run off a 3,1 logic board with 3,1 memory and all the other attending differences. Their performance suffers accordingly compared to a 4,1 or 5,1. Bottom line this affects single core performance, multi core performance, everything. A 3,1 will not fare well against a 5,1 with an upgraded CPU in any context. Enjoy the thin gruel of these benchmarks.
Believe what you like, but a Mac Pro 3,1 with 3.2GHz X5482 does NOT give 98% the performance of a Mac Pro 5,1 with 3.07GHz X5675. It's not even close.
Much depends on testing method. In some cases it will be close (where RAM speed does not matter that much), in other it won't.
Let's take the Cinebench R15 Single Core test: 5482 - 88 pts, 5690 - 111 pts. Hell, 5675 scores 82 pts in this test.
Let's take the Cinebench R15 Single Core test: 5482 - 88 pts, 5690 - 111 pts. Hell, 5675 scores 82 pts in this test.
I was referring to this one, because it lists OS version and machine type, as results may vary a little between platforms and OSes: http://cbscores.com/
I don't know I don't have a 5,1 but all the benchmarks you keep linking to show that the single threaded CPU performance of the 3.20GHz X5482 is pretty comparable to the Xeons in the 5,1 e.g.
X5670 97-99
X5482 88
i.e. the 5482 is 90% of the performance of the X5670 That looks pretty close to me. It's not like comparing them with the latest iMac where the 4.0GHz Core i7-6700K is double the speed of the X56xx processors.
It's difficult to put a quantitative value on it... There's nothing like sitting down and using both yourself to feel and judge for yourself how much of a difference there is. Personally, I've gone through a 1,1, a 3,1, and a 5,1. Going from 3,1 was a huge leap forward for me.
I haven't done much reading about the iMacs and haven't use one for more than a few minutes so I can't say how much truth there is to the "double the speed" claims.
I haven't done much reading about the iMacs and haven't use one for more than a few minutes so I can't say how much truth there is to the "double the speed" claims.