GeekBench Drops 31,000 to 18,000...

[lchenye]

So this is my situation, just bought a pre-owned mac pro 2009 from eBay( already sounds like a beginning of a tragedy). 3.46 12 core 128 ram with pcie ssd and titan x. On the item description there's a geek bench screenshot that shows 31000+ on august 2015. And when I received the mac today, just a year and a half, ran the geek bench, it becomes 18000+. SO! Is that possible that using a mac for a year and a half can lead to this kind of drop? Or I definitely ggot a lemon?

Will post the screenshots later, just couldn't sleep thinking I might be f**ked.

 

[Filin]

Geekbench change scores from version to version
18000 for your system is normal
see http://browser.primatelabs.com/v4/cpu/search?utf8=✓&q=x5690

 

[Plato65]

Were they done using the same version of Geekbench?

The score should be in the 18000-22000 range:
https://browser.primatelabs.com/v4/cpu/search?utf8=✓&q=platform:"Mac"+processor:"Intel+Xeon+X5690"

 

[h9826790]

As the other said, you have to use the exact same version of GB to benchmark your system. Otherwise, the result is not comparable.

Also, unless you can feel that your computer's real world performance do drop 40%, I suggest that you can simply forget about the benchmark. Especially if you have no idea what those score means, or how they determine the score.

I am not againsting any benchmark. However, you have to use them preoperly, they are just tools. Miss use a tool usually can't achieve your target and may even lead to damage.

 

[lchenye]

Okay I guess I just fliped out and doesn't pay attention to the detail. These two result is from different version of benchmark. I'll go check the actually data from the bench mark to see if there's huge different.

But what he did is, unless I have some major misunderstanding of benchmark - So in the description he put the 31k+ result with the 2013 mac pro which is around 23K+, to show that the tower is way better. Now I think the 23K+ result should be a different version of it?

I am not good with hardware but I go for the old tower because I think it have a more cost-effective than the trashcan( I don't really use thurdbolt a lot). So in this case, I paid like 3K for the old machine (12-core 3.46, 128G ram, pcie ssd with titan X ) with benchmark lower than the 2013 mac pro. Please tell me that benchmark is just a number...

If it helps, I plan on use the mac pro for post producing for film.

 

[h9826790]

Okay I guess I just fliped out and doesn't pay attention to the detail. These two result is from different version of benchmark. I'll go check the actually data from the bench mark to see if there's huge different.

But what he did is, unless I have some major misunderstanding of benchmark - So in the description he put the 31k+ result with the 2013 mac pro which is around 23K+, to show that the tower is way better. Now I think the 23K+ result should be a different version of it?

I am not good with hardware but I go for the old tower because I think it have a more cost-effective than the trashcan( I don't really use thurdbolt a lot). So in this case, I paid like 3K for the old machine (12-core 3.46, 128G ram, pcie ssd with titan X ) with benchmark lower than the 2013 mac pro. Please tell me that benchmark is just a number...

If it helps, I plan on use the mac pro for post producing for film.

GB4 was not avail back in 2015. However, most likely he was NOT posting the 12 cores nMP's benchmark result. The 12 cores nMP also score ~30000 in GB3. 23000 may be from a 8 cores cMP.

 

[DearthnVader]

Okay I guess I just fliped out and doesn't pay attention to the detail. These two result is from different version of benchmark. I'll go check the actually data from the bench mark to see if there's huge different.

But what he did is, unless I have some major misunderstanding of benchmark - So in the description he put the 31k+ result with the 2013 mac pro which is around 23K+, to show that the tower is way better. Now I think the 23K+ result should be a different version of it?

I am not good with hardware but I go for the old tower because I think it have a more cost-effective than the trashcan( I don't really use thurdbolt a lot). So in this case, I paid like 3K for the old machine (12-core 3.46, 128G ram, pcie ssd with titan X ) with benchmark lower than the 2013 mac pro. Please tell me that benchmark is just a number...

If it helps, I plan on use the mac pro for post producing for film.

Is the Titan X Flashed?

When we consider that you could have gotten a 6 Core dual D500 nMP for $3k, you may have over paid, but the nMP requires expensive Thunderbolt solutions to add Drives and PCIE cards. IF your needs are multiple drives, or you need something like and SDI card, the cPM is a no brainer.

Assuming your going to be using FCPX, you may want to try the BruceX benchmark, and compare it with the $3k nMP.

 

[Bollockser]

Is there a published list of the OLD geekbench scores, like the ones on Everymac.com from GB2 and GB3? I wish geekbench would stop squashing the top scores down with each new version, are they afraid of bigger benchmarks?

 

[ActionableMango]

Please tell me that benchmark is just a number...

It IS just a number and may or may not relate to what you are doing at all. You should look at benchmarks that run the actual application you use, like @DearthnVader's suggestion (assuming you use FCPx).

 

[lchenye]

GB4 was not avail back in 2015. However, most likely he was NOT posting the 12 cores nMP's benchmark result. The 12 cores nMP also score ~30000 in GB3. 23000 may be from a 8 cores cMP.

Oh I wasn't clear. I think when he made the eBay page which might be weeks before, he put ver3 GB of the tower and ver4 GB of the nMP together.
[doublepost=1492716037][/doublepost]

Is there a published list of the OLD geekbench scores, like the ones on Everymac.com from GB2 and GB3? I wish geekbench would stop squashing the top scores down with each new version, are they afraid of bigger benchmarks?

I think when you click into GB3 tab in their website, you can still search for the old GB marks, but not as friendly as GB4
[doublepost=1492716325][/doublepost]

Is the Titan X Flashed?

When we consider that you could have gotten a 6 Core dual D500 nMP for $3k, you may have over paid, but the nMP requires expensive Thunderbolt solutions to add Drives and PCIE cards. IF your needs are multiple drives, or you need something like and SDI card, the cPM is a no brainer.

Assuming your going to be using FCPX, you may want to try the BruceX benchmark, and compare it with the $3k nMP.

Yes clearly my poor hardware knowledge that leads me to believe 12 core 3.46 must be faster than the 6 core... But please tell me that the 128G ram and Titan X are not worse deal, also he got a 512 apple pcie SSD, RW about 980MB/s... I use mostly premiere and what worries me more, Davinci.
[doublepost=1492716439][/doublepost]

It IS just a number and may or may not relate to what you are doing at all. You should look at benchmarks that run the actual application you use, like @DearthnVader's suggestion (assuming you use FCPx).

You are right. I have never run Benchmark once in my life. It is just this time with eBay and pre-owned, I have to be clear if I got the right machine. I'll look for some benchmark for postproduction.

 

[flowrider]

This thread may interest the OP:

https://forums.macrumors.com/threads/the-enigma-called-geekbench.2039828/

Lou

 

[lchenye]

This thread may interest the OP:

https://forums.macrumors.com/threads/the-enigma-called-geekbench.2039828/

Lou

That's interesting...

What I've done is reset the PRAM and the RAM runs from 1066 to 1333. And the GB3 went up like a 1000 point, not a lot, but GB4 went up from 18l+ to 20K+ which is the similar score with 8-core nMP. Another interesting fact...

But anyway I lost my interests in bugging myself with this, I will just run some project on avid and DaVinci to see if it is fast enough for me.

Sorry but two more question,

1) the seller gave me a apple ssd SM0512f, and I have a choice for getting the sm951 which I believe it is better right? Even with the apple ssd is more expensive.

2) Now every time I start up the mac, it takes about 30 seconds before I hear the Chime, which I believes it is searching for boot drive. Tried work with the startup disk in preference, no help... I have no hard drives installed. I have only one PCIE ssd. (should I start a new thread about this?)

 

[h9826790]

SM951 (AHCI) is the best SSD you can get for cMP.

30s to chime because there are 128GB RAM to initialise / self test. If you remove most of the sticks, the boot time will be significantly faster.

More RAM has almost no down side in terms of performance consideration. Especially if you deal with multi GB videos. Then the RAM can serve as cache and keep lots of data in there, reduce the chance of loading the same data again from the hard drive. Which can actually speed up some process quite a bit (depends on your work flow).

TitanX is also a very good performer. Some users report keep hitting software bug, but most users report very happy with this card. So, most likely you will be OK with it.

 

[lchenye]

SM951 (AHCI) is the best SSD you can get for cMP.

30s to chime because there are 128GB RAM to initialise / self test. If you remove most of the sticks, the boot time will be significantly faster.

Thanks, I'll get the 951 then!

It's good to know that. I don't mind the boot time being longer, just want to make sure there isn't anything wrong about it... I'll do a test with 2 sticks to see if it is the problem

 

[Bollockser]

I knabbed a 512GB SM951 a couple weeks ago, $379 on eBay