Sidebar Sidebar
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

Overclocking Tool for the Mac Pro

from the aint-that-obvious-department

If this tool also allows for underclocking, I see some real applications. Seriously underclocking my Macbook Pro when I'm just surfing sounds like a good thing for both my battery and my lap.
 
No luck :(

Instant freeze here on my Mac Pro dual 2.8, and I barely moved to 405FSb to get started.

Probably ram related : got 4x2 gb OWC and 2x 1Gb original Apple Ram, still not even 1% extra kind of bugs me :confused:
 
diamond.g said:
Is there no way to lock the memory bus speed down? I know on some motherboards you can choose to run the memory at 1:1, 5:4 or some other variation. That would minimize the impact of upping the bus speed on the RAM. It also happened to be what I was thinking about locking down, but for some reason forgot to mention it. Besides I am not sure how else people would be hitting higher clock speeds as memory stability tends to hold everyone back.
Click to expand...

Yes, the FSB:DRAM ratio can be set to 5:4. However, this must be set before the OS loaded. If you change that afterwards, the FB-DIMMs will loose data. But I can play around with this. Maybe I'll find method, how to smoothly change the FSB:DRAM ratio by changing some other parameter of the MCH (RAS, CAS, etc.)

-Christoph
 
gnasher729 said:
I'd say you are blaming Apple for something that is not their fault at all.

On AMD processors, the rdtsc instruction has real problems because the timers for multiple CPUs drift apart, making it very hard indeed to use rdtsc for any clock purposes (some linux guys probably know all the details). Intel CPUs don't have that problem. And since Apple doesn't use AMD CPUs, that AMD problem is irrelevant. If an AMD Hackintosh has problems because of that, tough. Not Apple's fault.

The rdtsc assembler instruction itself has the problem that it counts motherboard clocks (multiplied by the maximum multiplier of the processor) and doesn't directly give a time, so people usually use mach_absolute_time. Apart from that, rdtsc is _not_ influenced by what you call "underclocking" in a notebook, which is really reducing the clock multiplier. In every second, rdtsc is increased by a number equal to the maximum clock speed of the processor, no matter what the actual clock speed is.

mach_absolute_time, on the other hand, returns time in nanoseconds, and it does that correctly on multi-CPU systems: First, it uses rdtsc to read the current clock. Then it subtracts a correction factor that can be different per CPU. Then it multiplies by a factor to change clocks to nanoseconds. It then adds another correction factor, and finally checks whether it is still running on the same CPU it was running on when the call started - if not, then it does the whole calculation again.

Here is the point where this overclocking tool is missing something: As it changes the motherboard clock speed and therefore the rate at which rdtsc is counting, it should change the multiplier used by mach_absolute_time and change the correction factors that are subtracted and added so that mach_absolute_time can continue to give the correct results. It doesn't do that. Therefore mach_absolute_time will give incorrect results. There is no way you can blame Apple for that - I am sure that if Apple provided an overclocking tool then they would make this work, and the stupid rebooting would probably not be needed.
Click to expand...

"Underclocking" on notebooks is not only done by lowering the multiplier. On Centrino 2 chipsets, the bus clock can also be reduced. Windows does this for instance, don't know, if Mac OS X uses this technique.

If a (mostly notebook) processor still gets too hot, it throttles down by simply leaving out bus cycles.

This leads to a deviation of real time and bus time.

Using the bus clock as real time source is simply bad bad practice. RDTSC always returns bus clock dependent time. Whereas RDTSC is a CPU instruction, mach_absolute_time is a system call. This gives reliable only results, when implemented correctly, i.e. derived from a real time source.

-Christoph
 
Vielen Dank fürs Dienstprogramm, Christoph!!!

Restarting cures the system clock problem, but what if you have one of those MacPros where the SuperDrive locks up, going whirr-chick, whirr-chick, whirr-chick, when you restart? Anyone found a solution for that?
 
christoh said:
Yes, the FSB:DRAM ratio can be set to 5:4. However, this must be set before the OS loaded. If you change that afterwards, the FB-DIMMs will loose data. But I can play around with this. Maybe I'll find method, how to smoothly change the FSB:DRAM ratio by changing some other parameter of the MCH (RAS, CAS, etc.)

-Christoph
Click to expand...

Sweet!, good luck.
 
christoh said:
Using the bus clock as real time source is simply bad bad practice. RDTSC always returns bus clock dependent time. Whereas RDTSC is a CPU instruction, mach_absolute_time is a system call. This gives reliable only results, when implemented correctly, i.e. derived from a real time source.
Click to expand...

It gives reliable results when it is based on rdtsc, and whoever plays around with the motherboard clock can be bothered to fix a total of _three_ values stored in fixed locations that are used to translate clocks into nanoseconds.
 
gnasher729 said:
It gives reliable results when it is based on rdtsc, and whoever plays around with the motherboard clock can be bothered to fix a total of _three_ values stored in fixed locations that are used to translate clocks into nanoseconds.
Click to expand...

Sounds great, are these values exported to any library, e.g. com.apple.kpi.unsupported. We appreciate any hints, that help to make ZDNet Clock better.

-Christoph
 
gatd said:
what speed can a 3.2 mac pro be clocked to?
Click to expand...

That really depends on the system and the memory. Every computer has different OC capabilities. You will have to test it out yourself to find how high you can go.

Oddly a general rule of thumb is the lower clocked chips usually can run as fast as the highest sold chip (3.2Ghz I believe). The higher clocked chips can run faster but the OC percentage won't be as high (before having to up the voltage, usually).
 
Ok, so is this the kind of thing that one should avoid while one's system is still covered under Applecare? It occurs to me that if one had any/some/most of the symptoms outlined as 'disadvantages' and Apple found that you'd been overclocking your hardware, it might be their contention that you drove it to the problems for which you're asking for assistance under Applecare... Or am I over-thinking?:cool:

Thanx.
 
The Marketing Bus

Was really intrigued about the 'marketing bus' (great name by the way!) being 400Mhz and Intel say your bus speed is 4 x 400Mhz i.e. 1600Mhz FSB
Is this 400Mhz per core?

...or maybe this is a wacky bus all those crazy marketing guys at Apple get on to go to work!
 
BongoBanger said:
The Mac Pro uses Xeon chips running at 80w. The iMac uses modified Centrino chips running at 55w. In other words the iMac chip has already been boosted but uses less power and therefore generates less heat. In saying that, 55w is pretty heavy for what is essentially a notebook CPU. The Mac Pro obviously has much better cooling (because of case size and fan size) so you can do more with it.
Click to expand...

A couple of these monsters help too :) --->

http://powermax.download-files.net/macpro2/processor_sinc.jpg
 
Schizoid said:
Was really intrigued about the 'marketing bus' (great name by the way!) being 400Mhz and Intel say your bus speed is 4 x 400Mhz i.e. 1600Mhz FSB
Is this 400Mhz per core?

...or maybe this is a wacky bus all those crazy marketing guys at Apple get on to go to work!
Click to expand...

Each core shares the same bus, so no it isn't 400Mhz per core. Intel has been doing the 'quad pumped bus' since the P4 days. When they go IME it won't be 'quad pumped'.
 
brent1a said:
From reading the other posts I'd say the ram is the problem. I have an extra 2gb's of iRam from Newegg but I haven't had any problems so far.

8800 gfx card, huh? I'm jonesing for the ATI 3870 real bad, myself.
:)
Click to expand...

Thnx for input. I wonder of this has actually done me a favor. If my extra 4 GIGs of RAM (guaranteed for life) cannot support more than 6% does this mean it is not too good and may be causing other problems, or is it simply incompatibility with Apple's own installed 2 GIGs?

Yes NVIDIA GeForce 8800 GT is nice but having the stock ATI Radeon HD 2600 XT in too has given me a few problems. Pro apps can get confused and freeze up if I move windows from one screen to the other. I can't work without 2 ACD screens I may have to add a second NVIDIA GeForce 8800 GT and remove the 2600.
 
Hi Christoh, does my inability to exceed 6% indicate poor 3rd party RAM or simply incompatibility between Apple's stock 2 GIGs and my additional 4 GIGs? If the former I am glad I found out! Your utility may have additional uses ;)

It was supposed to be high quality with lifetime guarantee ... :(

I do now see it has no serial numbers whereas Apple RAM does which makes me feel a little uneasy.

Here is the info on the 2 2 GIGs.

DIMM Riser B/DIMM 2:

Size: 2 GB
Type: DDR2 FB-DIMM
Speed: 800 MHz
Status: OK
Manufacturer: 0x05F7
Part Number: 0x000000003732353642363145353636374600
Serial Number: 0x00000000

DIMM Riser B/DIMM 1:

Size: 2 GB
Type: DDR2 FB-DIMM
Speed: 800 MHz
Status: OK
Manufacturer: 0x05F7
Part Number: 0x000000003732353642363145353636374600
Serial Number: 0x00000000
 
Digitalclips said:
Hi Christoh, does my inability to exceed 6% indicate poor 3rd party RAM or simply incompatibility between Apple's stock 2 GIGs and my additional 4 GIGs? If the former I am glad I found out! Your utility may have additional uses ;)

It was supposed to be high quality with lifetime guarantee ... :(

I do now see it has no serial numbers whereas Apple RAM does which makes me feel a little uneasy.

Here is the info on the 2 2 GIGs.

DIMM Riser B/DIMM 2:

Size: 2 GB
Type: DDR2 FB-DIMM
Speed: 800 MHz
Status: OK
Manufacturer: 0x05F7
Part Number: 0x000000003732353642363145353636374600
Serial Number: 0x00000000

DIMM Riser B/DIMM 1:

Size: 2 GB
Type: DDR2 FB-DIMM
Speed: 800 MHz
Status: OK
Manufacturer: 0x05F7
Part Number: 0x000000003732353642363145353636374600
Serial Number: 0x00000000
Click to expand...
A 6% OC? For RAM spec'ed @ 800 Mhz that isn't bad at all. I wouldn't expect all RAM to run at higher than spec'ed rates.
 
Digitalclips said:
Thnx for input. I wonder of this has actually done me a favor. If my extra 4 GIGs of RAM (guaranteed for life) cannot support more than 6% does this mean it is not too good and may be causing other problems, or is it simply incompatibility with Apple's own installed 2 GIGs?
Click to expand...

You could try swapping RAM chips around.

There will be tiny differences in timing between the different slots. Apple's engineers have made a design where a signal takes X nanoseconds to receive the memory chip, and the chip then is allowed to take Y nanoseconds to respond. Apple and the RAM maker guarantee that they meet the requirements (X and Y nanoseconds). But in reality they will work a bit faster, which is what allows you to run 6% faster, and they also take slightly different amounts of time. Maybe you had the slowest RAM chips in the slowest slot; by swapping around you might get a combination that works better.

Especially the front and back slots in each bank are different; swapping front and back slots is reasonable likely to run either slower or faster.
 
Schizoid said:
Was really intrigued about the 'marketing bus' (great name by the way!) being 400Mhz and Intel say your bus speed is 4 x 400Mhz i.e. 1600Mhz FSB
Is this 400Mhz per core?

...or maybe this is a wacky bus all those crazy marketing guys at Apple get on to go to work!
Click to expand...

Remember Pentium 4 days? Intel built processors that would run at 3,000 MHz without getting much work done and people built them like crazy because 3000 MHz must be faster than 2000 MHz. AMD did strike back by building a 2,000 MHz chip and calling it "AMD 3000" or something like that :D

Years ago there was the "scanner resolution" race. Some scanners shipped that produced 300dpi real resolution and 9600 dpi "interpolated resolution". Which means for every real pixel the scanner added 31 more pixels horizontally and 31 more pixels vertically; the only effect was using 1000 times more storage for no quality gain whatsoever. But the scanners sold because they were 9600 dpi instead of 300 dpi.

Ever used a video camera with a "50 times digital zoom"? The result: 10 pixels blown up to the size of your full screen. You can't even guess what's shown on the picture. :mad:
 
diamond.g said:
A 6% OC? For RAM spec'ed @ 800 Mhz that isn't bad at all. I wouldn't expect all RAM to run at higher than spec'ed rates.
Click to expand...

You missed my point though, others are getting way higher and while I might use ZDNet Clock or not, currently I am more concerned that it might be showing up a bad RAM configuration I was unaware of, in which case I am really happy I tried it.
 
gnasher729 said:
You could try swapping RAM chips around.

There will be tiny differences in timing between the different slots. Apple's engineers have made a design where a signal takes X nanoseconds to receive the memory chip, and the chip then is allowed to take Y nanoseconds to respond. Apple and the RAM maker guarantee that they meet the requirements (X and Y nanoseconds). But in reality they will work a bit faster, which is what allows you to run 6% faster, and they also take slightly different amounts of time. Maybe you had the slowest RAM chips in the slowest slot; by swapping around you might get a combination that works better.

Especially the front and back slots in each bank are different; swapping front and back slots is reasonable likely to run either slower or faster.
Click to expand...

Ok, help me out here ... sorry to be DIMM ;)
This Mac Pro is the first Mac I have seen wher the RAM is installed this way.

On A DIMM Riser is both Apple RAM DIMMs in DIMM slots 1 & 2.
On B DIMM Riser is both 3rd Party RAM DIMMs in slots 1 & 2 also

Where are you suggesting I move what to...? Do you mean switch 3rd party positions in slots 1 & 2 or move to 3 & 4 or are you suggesting switching Risers?
 
Meh...if Apple fixed the MacPro firmware problems (I think mine is first gen .. 2x 2.33GHz Dual...not sure) then you can get a 10 to 20% speed up with a few key presses.

Try it, restart and do a PRAM reset.....feel how much snappier your system is....do an XBench and save the results.

Now, either let the system sleep or restart it again with no PRAM zap....machine slows down, do an XBench and compare the results.....

....there's a 15/20% performance difference between the two scenarios on my machine.

Seems like there's a firmware bug where it doesn't save settings correctly or something.

I reported this to Apple on two seperate occasions, back when my MacPro still had AppleCare coverage....not even a response.
 
You must log in or register to reply here.
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.
Top Bottom
Back