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.
Overclock dual 1.42 MDD... performance gain??
- Thread starter Nova77
- Start date
- Sort by reaction score
A fast, modern HDD can make a substantial difference in performance. New SATA or eSATA setups can move data quite a bit faster than the onboard ATA system.
....... where are the infos, comparison of before and after the overclock... and well anything that answers my question??... please...
----------
I said, overclocking or other solution than hard drive and ram.
I can't go much further with RAM and HDD upgrade anyways. I've passed the 100 MB/sec mark, so... you know... Maybe SATA 2 card, but **** those can't boot and costs 3X the price.
You can try overclocking the bus. But that is a very touchy affair, and I'm not sure how well the MDD responds to a system bus overclock.
thats a good question... anybody knows if it is possible to overclock the FSB to more than the stock 167mhz?? I know theres a trick for the 133mhz MDDs, but that is NOT the question here...
Exactly what you would expect mathematically. It's linear.
The difference between my 1.42 and 1.5 with no other changes to system:
1.42:
Results 66.64
System Info
Xbench Version 1.3
System Version 10.5.8 (9L31a)
Physical RAM 2048 MB
Model PowerMac3,6
Processor PowerPC G4x2 @ 1.42 GHz
Version 7455 (Apollo) v3.3
L1 Cache 32K (instruction), 32K (data)
L2 Cache 256K @ 1.42 GHz
L3 Cache 2048K @ 237 MHz
Bus Frequency 167 MHz
Video Card ATY,R350
Drive Type ST3500418AS ST3500418AS
CPU Test 76.67
GCD Loop 135.60 7.15 Mops/sec
Floating Point Basic 49.63 1.18 Gflop/sec
AltiVec Basic 279.53 11.14 Gflop/sec
vecLib FFT 83.18 2.74 Gflop/sec
Floating Point Library 45.27 7.88 Mops/sec
Thread Test 74.53
Computation 110.92 2.25 Mops/sec, 4 threads
Lock Contention 56.12 2.41 Mlocks/sec, 4 threads
Memory Test 49.94
System 57.88
Allocate 219.72 806.90 Kalloc/sec
Fill 56.66 2754.73 MB/sec
Copy 33.75 697.10 MB/sec
Stream 43.92
Copy 46.78 966.27 MB/sec [altivec]
Scale 47.49 981.05 MB/sec [altivec]
Add 43.27 921.66 MB/sec [altivec]
Triad 39.17 837.89 MB/sec [altivec]
Quartz Graphics Test 86.61
Line 73.60 4.90 Klines/sec [50% alpha]
Rectangle 78.24 23.36 Krects/sec [50% alpha]
Circle 70.59 5.75 Kcircles/sec [50% alpha]
Bezier 81.44 2.05 Kbeziers/sec [50% alpha]
Text 203.25 12.71 Kchars/sec
OpenGL Graphics Test 76.31
Spinning Squares 76.31 96.81 frames/sec
User Interface Test 51.33
Elements 51.33 235.59 refresh/sec
Disk Test 69.28
Sequential 128.44
Uncached Write 144.58 88.77 MB/sec [4K blocks]
Uncached Write 142.07 80.39 MB/sec [256K blocks]
Uncached Read 91.10 26.66 MB/sec [4K blocks]
Uncached Read 161.03 80.93 MB/sec [256K blocks]
Random 47.43
Uncached Write 15.70 1.66 MB/sec [4K blocks]
Uncached Write 222.19 71.13 MB/sec [256K blocks]
Uncached Read 101.71 0.72 MB/sec [4K blocks]
Uncached Read 158.33 29.38 MB/sec [256K blocks]
1.5:
Results 71.72
System Info
Xbench Version 1.3
System Version 10.5.8 (9L31a)
Physical RAM 2048 MB
Model PowerMac3,6
Processor PowerPC G4x2 @ 1.50 GHz
Version 7455 (Apollo) v3.3
L1 Cache 32K (instruction), 32K (data)
L2 Cache 256K @ 1.50 GHz
L3 Cache 2048K @ 250 MHz
Bus Frequency 167 MHz
Video Card ATY,R350
Drive Type ST3500418AS ST3500418AS
CPU Test 81.35
GCD Loop 143.74 7.58 Mops/sec
Floating Point Basic 52.66 1.25 Gflop/sec
AltiVec Basic 297.59 11.86 Gflop/sec
vecLib FFT 88.32 2.91 Gflop/sec
Floating Point Library 47.99 8.36 Mops/sec
Thread Test 88.62
Computation 121.92 2.47 Mops/sec, 4 threads
Lock Contention 69.61 2.99 Mlocks/sec, 4 threads
Memory Test 51.83
System 64.44
Allocate 277.47 1.02 Malloc/sec
Fill 64.32 3127.43 MB/sec
Copy 36.49 753.63 MB/sec
Stream 43.35
Copy 45.42 938.19 MB/sec [altivec]
Scale 47.27 976.53 MB/sec [altivec]
Add 41.99 894.47 MB/sec [altivec]
Triad 39.54 845.87 MB/sec [altivec]
Quartz Graphics Test 93.12
Line 75.04 5.00 Klines/sec [50% alpha]
Rectangle 83.83 25.03 Krects/sec [50% alpha]
Circle 80.02 6.52 Kcircles/sec [50% alpha]
Bezier 88.64 2.24 Kbeziers/sec [50% alpha]
Text 214.66 13.43 Kchars/sec
OpenGL Graphics Test 80.02
Spinning Squares 80.02 101.51 frames/sec
User Interface Test 58.33
Elements 58.33 267.70 refresh/sec
Disk Test 69.69
Sequential 128.73
Uncached Write 145.26 89.19 MB/sec [4K blocks]
Uncached Write 142.69 80.74 MB/sec [256K blocks]
Uncached Read 90.53 26.49 MB/sec [4K blocks]
Uncached Read 163.01 81.93 MB/sec [256K blocks]
Random 47.77
Uncached Write 15.77 1.67 MB/sec [4K blocks]
Uncached Write 228.73 73.23 MB/sec [256K blocks]
Uncached Read 103.94 0.74 MB/sec [4K blocks]
Uncached Read 157.73 29.27 MB/sec [256K blocks]
1.58 was also stable for mine, but the temperature jumped a little above my comfort level so I dropped it back down (my comfort level for temperature is much below manufacturer's specs).
Of course, not all 1.42's are going to be the same. As long as you're comfortable with your soldering skills, though, it doesn't hurt try for something higher and then drop it back down if it doesn't work out. I think 1.5 is pretty much a given for any 1.42. My impression based on what I've read and done (I can't quantify this in any way. People are probably more likely to post about their victories than their failures, so I'm basically pulling this out of my you know what) is that 1.5 is a given, 1.58 works more often than not, and the odds are against you at 1.67 (people seem to have more success with water cooling). If I'm not mistaken, though, at least one user here has gotten 1.67 without water cooling. Maybe 666Sheep?
I would strongly urge you, Nova77, to consider using CHUD if this is the next step for your MDD. It might make the difference between something you're comfortable with and something that you're not. The cooler they run, the longer they will last.
@ThunderSnake thanks a lot, that is the kind of infos I was looking for.
Now will 666sheep get out of the shadows and share with us his knowledge? lol
Heh. I just called him out in another thread, so maybe he'll show up.
Totally agree. To me, it's not worth it if you're going to be pushing these chips to the limits of their design specification. Keep an eye on those temps. I have another 1.42 daughtercard that I haven't messed with yet that I might try to get 1.58 out of, but not if the temps climb to the high 50's on a regular basis. My 1.5 rarely gets above 50C. The fans never kick into high speed. It's very quiet. Again, CHUD helps a lot.
Overclocking is a fun hobby, but the PowerPC CPUs in Macs don't respond all that well to overclocking (especially when you compare them with x86 CPUs).
Personally I'd leave the dual 1.42 MDD alone in terms of its CPUs.
Personally I'd leave the dual 1.42 MDD alone in terms of its CPUs.
I never overclocked my 2 dual MDD's (867 and 1.42) but I did run nap and replaced all the stock fans. Doing lower to mid level taxing of the systems would keep the 867 at 42-43c and 1.42 at 46-48. Because I added an external 120mm on both to help exhaust heat they both never went above maybe 52 even under 100% load for hours.
Me too. Sort of. I added an 80mm exhaust fan internally between the CPU and the case. It's mounted high enough that it helps vent GPU heat too. Great minds think alike!
Everybody knows what overclocking does... yep more heat... But this is totally off topic guys, I was asking about performance, not if I should do it or not in your opinion.
Btw heat is not so bad... my MDD runs AT MOST @ 56 °C in hot summer, and it is not a genuine 1.42 MDD so the CPU fan is less powerful. I have no cooling mods at all. I don't know why everyone seems to have such bad heat issues with MDDs.... Maybe its time for them to replace that old thermal paste??!!
Just by changing the fan and adding 2 small fans near the CPU (you know, what everybody does just for cooling their non-overclocked MDD), I think it would easily reach 1.5 ghz without major heat issues.
Oh well, I did continue off track....
Now, back on topic...!
Btw heat is not so bad... my MDD runs AT MOST @ 56 °C in hot summer, and it is not a genuine 1.42 MDD so the CPU fan is less powerful. I have no cooling mods at all. I don't know why everyone seems to have such bad heat issues with MDDs.... Maybe its time for them to replace that old thermal paste??!!
Just by changing the fan and adding 2 small fans near the CPU (you know, what everybody does just for cooling their non-overclocked MDD), I think it would easily reach 1.5 ghz without major heat issues.
Oh well, I did continue off track....
Now, back on topic...!
I get that you understand the risks, but lots of people read these things, so I think it's appropriate to include some cautions. The only opinion that I've offered you specifically is that I think you should run CHUD.
Hmmm.... no?... Is that your sense of humor zen.state?
Don't know if I sound like a boss or something but thats not the goal... XD
It was just a reminder of what this thread is about :
Performance.
It's not about heat. Thats it. Reply the way you want, but if you want to start a new topic, why not start another thread...
So this is the best place to post benchmarks of before/after overclocking, just like Thundersnake did.
Last edited:
Exactly, esp. in synthetic benchmarks.
TS: You did ask about WC in another thread: no, mine is aircooled (still... I've stuck on case surgery stage + looking for waterblock for 7800GS). But it's way too loud (2x Delta 1212SHE, I got rid of optical drive)
I had to raise CV a lot from stock value and without massive fans CPU and L3 chips are overheating very fast. WC is on it's way Let's see what bizzle will be able to reach. Each daughtercard may have a different OC potential, but I'm 99% sure that 1.67 is max for 7455 w L3. Question is: with what Core Voltage
Exactly, esp. in synthetic benchmarks.
TS: You did ask about WC in another thread: no, mine is aircooled (still... I've stuck on case surgery stage + looking for waterblock for 7800GS). But it's way too loud (2x Delta 1212SHE, I got rid of optical drive)
Let's see what bizzle will be able to reach. Each daughtercard may have a different OC potential, but I'm 99% sure that 1.67 is max for 7455 w L3. Question is: with what Core Voltage
Can you post some benchmarks of your 1.67 MDD? Or other infos.... I want to know if its worth the overclock.
http://db.xbench.com/merge.xhtml?doc1=359219
https://forums.macrumors.com/posts/10637338/
Geekbench seems too high, something strange happens in text decompress, but every time it acts like this...
https://forums.macrumors.com/posts/10637338/
Geekbench seems too high, something strange happens in text decompress, but every time it acts like this...
Are you certain that it was the CPU? What temperature did it run? At idle? At full load? Were you running CHUD? Did you apply fresh thermal compound? (Please don't be insulted by that last question, but I'm surprised by how many people do not know to do this. I've found that it's worth asking even with folks savvy enough to do their own overclocking.) Anything additional that you have to offer could be helpful.
Only about 18 months on my MDD, but I've had an OC'd DA module running fine for about seven years or so. As long as they run cool (of course, "cool" is a relative term when talking MDDs), they should be fine.