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

2011 Mac Mini Server RAM @ 1867MHz

S

shortcut3d

macrumors 65816
Original poster
Aug 24, 2011
1,112
15
I just received 8GB (4GBx2) Kingston HyperX PnP 1867MHz RAM and began testing.

Under Windows 7 x64 the memory experience score is a perfect 7.9 :D

Geekbench 64-bit score is 9762 compared to the iMac in my signature with 32GB of RAM Geekbench score of 10712.

I'm running Rember to check stability.

Geekbench, Rember and System Profiler all report it's at 1867MHz.
 
the ram needs to be tested in windows by playing a game that pushes the intel graphics hard. it may be that you can boost the graphics by as much as 30-40 percent. this would make the server even better then anyone ever thought it would be when playing games. if you make one hdd for windows and one for mac you could have an amazing machine.

put windows on one hdd and mac on the other test a game on each hdd osx then pull the 1867 ram and put the 1333 ram in.. it would be nice if you get a big difference in performance.
 
I question if your memory is really running at 1876 MGhz. I suspect that System Profiler is simply reading the the speed the modules are rated at not the speed they are actually running. Geekbench is just pulling its information from System Profiler. Can't comment on Rember as I've never used it.

There are two 2011 Mac Mini Servers on Geekbench browser, both have 8GB of RAM running at 1333, one scored 9635 the other 9674. The performance boost your seeing over the base MM sever is due to the increased amount of RAM, not from a higher speed.

http://browse.geekbench.ca/geekbench2/search?q=mac+mini+server
 
Last edited:
Just read this on NewEggs site

FASTER JUST GOT EASIER
Kingston's HyperX PnP memory is programmed with faster frequencies and when 'plugged' into a system using the Sandy Bridge chipset, will automatically 'play' at rated speed in both desktop and notebook PCs - no further BIOS settings required. (Note: PnP implementation is only possible in configurations that include a BIOS that supports the PnP function. Your maximum speed will be determined by your BIOS.)
Click to expand...
Which would seem to imply that the memory could be running at 1867, the following casts some doubt on that though
Note: PnP implementation is only possible in configurations that include a BIOS that supports the PnP function. Your maximum speed will be determined by your BIOS.
Click to expand...
Apple computers use EFI rather than a BIOS, and the Apple EFI is locked down, no user options that I know of. Even if the memory is running at 1867 it doesn't seem to be giving you any better performance than you would have if you installed 8GB of 1333.
 
The MBP forum has confirmed that the memory is indeed running at a higher speed. The problem appears that the 2.0 i7 Quad-core may only be able to take advantage of 1600MHz. Some believe that the 2.2 GHz i7 or 2.3 GHz i7 would be required for 1866 MHz. Although the RAM is actually running at 1866 MHz.

My Geekbench scores may be off. I can re-run before I blow out the OSX partition. The Mini's are purely Windows 7 machine.

The Windows 7 index does seem to point to the memory being clocked higher an increase from 7.2 to 7.9 (perfect score). The graphics card took a minor bump from 6.2 to 6.3 but I was running 16GB prior and Windows 7 appears to allocate 1.5GB to the video card with 16GB.

I might try a game under Windows this weekend.
 
shortcut3d said:
The MBP forum has confirmed that the memory is indeed running at a higher speed.

I might try a game under Windows this weekend.
Click to expand...

I'm curious as to how they verified RAM speed. Would you be able to post a link to the MBP discussion, I did a search but couldn't find anything.

If you get a chance to post the game results, please do so. I'm interested in seeing what kind of performance your getting. My MMserver and 8GB of 1333 RAM should show up tomorrow. I'll try to run e few benchmarks this weekend and see how we compare.
 
Haifisch86 said:
That's very interesting! I didn't know PC memory worked in Macs...:confused:
Click to expand...

PCs and Macs use the same RAM. Apple branded RAM is just manufacturers taking advantage of people not knowing that they're the same and charging more.

I am surprised though that 1867MHz works in the new mini. As per Intel, it's only supposed to support 1333MHz. Interesting.
 
Intel's documentation is poor and changes between revisions. There is also the processor's access speed to DMI vs memory's. The benefit of faster memory would be limited. However, Windows experience score went to a perfect 7.9 and there was a little boost in the IGP score as well. This indicates the memory is actually at speed. There are still questions around the real world benefit / gain.

I'm going to try it in a dual core i7 mac mini in a few days. My guess is the PnP will set it to 1333MHz.

Geekbench scores tend to weigh more on capacity vs speed.
 
alust2013 said:
PCs and Macs use the same RAM. Apple branded RAM is just manufacturers taking advantage of people not knowing that they're the same and charging more.

I am surprised though that 1867MHz works in the new mini. As per Intel, it's only supposed to support 1333MHz. Interesting.
Click to expand...

I found documents on the server cpu and on the intel graphics that suggest 1600 will work. If he is getting 1867ghz it would make the server a very nice piece of gear.
 
philipma1957 said:
I found documents on the server cpu and on the intel graphics that suggest 1600 will work. If he is getting 1867ghz it would make the server a very nice piece of gear.
Click to expand...

I just saw a chart that showed supported RAM speeds, and for whatever reason, the minimum CPU to support 1600 was the 2.2 i7 quad. Clearly, that was incorrect.
 
gotzero said:
Thanks for the info! I just got a 2011 mini server and am looking for RAM. I guess I will get the faster clock.
Click to expand...

gotzero said:
Thanks for the info! I just got a 2011 mini server and am looking for RAM. I guess I will get the faster clock.
Click to expand...


To be honest I'm still leaning towards the point of view that the performance gain shortcut is seeing has more to do with upgrading from 4GB to 8GB than it does from increasing the RAM speed from 1333 to 1837 the performance gain shortcut. While there may be some slight additional gain from the memory speed keep in mind that performance is coming at a relatively high cost. The 8 GB of 1837 memory that allowed shortcut to hit 9762 in Geekbench cost $129, 8 GB of 1333 which is hitting 9672 in Geekbench costs about $45. In other words a 1% increase in Geekbench performance is costing you nearly three times as much.

The increase in 90 Geekpoints that shortcut saw may not even be indicative of an actual performance gain, and may in fact be due to variations brtween GB 2.2 which is the version shortcut used and 2.1.13 which is the version the Mac mini's with 8GB of 1333 was run.


From the GB web site -
One thing about the results, though, is that they're from different versions of Geekbench (2.1.13 vs 2.2.1). There are some changes in the individual tests from Geekbench 2.1 to Geekbench 2.2 that might cause the discrepancy you're seeing, so in the meantime I'd recommend running the same version of Geekbench on both machines or finding results from the same version of Geekbench.
 
I ran test under Windows 7 x64:

64 bit 11622
32 bit 8654

Apparently Windows benches higher?
 
Wirelessly posted (Mozilla/5.0 (iPhone; U; CPU iPhone OS 4_3_5 like Mac OS X; en-us) AppleWebKit/533.17.9 (KHTML, like Gecko) Version/5.0.2 Mobile/8L1 Safari/6533.18.5)

That's an interesting observation and something I hadn't really thought of. I guess it's possible. I'll take a look on GBs FAQ and see if Win vs OSX results are Apples to Apples or Oranges to Apples.
 
shortcut3d said:
I ran test under Windows 7 x64:

64 bit 11622
32 bit 8654

Apparently Windows benches higher?
Click to expand...

shhhh don't let anyone know how good it is with windows 7! I am coming to the conclusion that these machines run windows 7 better then they run lion. too bad they won't let snow leopard run on them correctly. as I think snow is better then lion.
 
philipma1957 said:
shhhh don't let anyone know how good it is with windows 7! I am coming to the conclusion that these machines run windows 7 better then they run lion. too bad they won't let snow leopard run on them correctly. as I think snow is better then lion.
Click to expand...

With AHCI enabled I get 7.4/7.9/6.3/6.3/7.9 Windows performance index score. Pretty insane results for a little box.
 
Nice to see they work. I'm looking to upgrade my 2.5 i5 mini to 8 Gb.
One question, how hot do they get?. HyperX memory has a reputation of running pretty hot.
Do you hear that the fan is running louder?.
Some users has been experiencing notebook hangs and resets due to higher temps using this modules.

Sorry for my english.

Cheers.
 
So far so good. I ran a stress test using Geekbench without issue. I also ran Rember 4 pass which is a few hours and passed. The fans never got so loud they were annoying, but they were noticeable. One machine has been up 24/7 since Tuesday. The second has been up and down due to Windows Media Center tweaking with HDHomerRun prime.
 
reputationZed said:
The performance boost your seeing over the base MM sever is due to the increased amount of RAM, not from a higher speed.
Click to expand...

I did some digging around on the Geekbench support forum and came across these replies from John at Primate Labs.

Generally speaking, the amount of RAM installed in your Mac doesn't affect the outcome of the tests. However, other details about the RAM (such as whether it's paired, or what its speed is) can have an effect on the results.
Click to expand...


Adding more memory to your computer only improves performance if your computer didn't have enough memory for your applications in the first place.

Testing for insufficient memory is difficult and highly dependent on each user's workflow. Some users only need 2GB of RAM while other users need 16GB of RAM. Instead, Geekbench tests the speed of the memory installed (i.e., how fast the RAM is at providing the processor with data). If your new RAM is the same kind as your old RAM (e.g., DDR2-800) your memory won't be any faster (or slower!) and thus your score won't change.
Click to expand...

So it looks like my assessment that GB scores had more to do with the amount of memory than they do with memory speed was off base.

Just finished running GB after upgrading my 2011 Mac Mini Server from 4GB of RAM 1333MHz to 8GB 1333MHz RAM.

64 bit test average of 5 runs for each set of RAM

4GB - 9379
8GB - 9577

so while it looks like doubling the RAM from 4GB to 8GB does increase GB scores some what the increase amount of RAM alone does not account for the performance scores shortcut3d was seeing with the 1866 MHz RAM.
 
Last edited:
Wirelessly posted (Mozilla/5.0 (iPhone; U; CPU iPhone OS 4_3_5 like Mac OS X; en-us) AppleWebKit/533.17.9 (KHTML, like Gecko) Version/5.0.2 Mobile/8L1 Safari/6533.18.5)

I also found the same quotes in the Geekbench forums.

In addition, the overall score is weighted so it may not capture the full benefit of the RAM speed increase. I'll look at my browser history for OSX memory scores, but I think I only saved Windows which are significantly higher than Lion.
 
Sorry for reviving an old thread but it is somewhat relevant to my issue. I noticed that some 2011 Mac Minis recognize Kingston HyperX 1867 Mhz RAM at 1867 Mhz. My 2011 Mini only recognizes this RAM at 1600 Mhz.

Does anyone know how I can tweak my Mini to recognize the RAM at 1867?
 
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