LinuxQuestions.org
Share your knowledge at the LQ Wiki.
Go Back   LinuxQuestions.org > Forums > Linux Forums > Linux - Hardware
User Name
Password
Linux - Hardware This forum is for Hardware issues.
Having trouble installing a piece of hardware? Want to know if that peripheral is compatible with Linux?

Notices


Reply
  Search this Thread
Old 11-07-2012, 12:59 PM   #46
Ulysses_
Member
 
Registered: Jul 2009
Posts: 756

Original Poster
Rep: Reputation: 43

Anyway, the i7-3770K is only slightly faster than the i7-3770T but much more energy hungry. So I took the risk and ordered an i7-3770T from a German shop, fingers crossed.

Is any DDR3-1600 memory at 8GB better than others?

Last edited by Ulysses_; 11-07-2012 at 01:02 PM.
 
Old 11-07-2012, 01:44 PM   #47
TobiSGD
Moderator
 
Registered: Dec 2009
Location: Germany
Distribution: Whatever fits the task best
Posts: 16,511
Blog Entries: 2

Rep: Reputation: 4546Reputation: 4546Reputation: 4546Reputation: 4546Reputation: 4546Reputation: 4546Reputation: 4546Reputation: 4546Reputation: 4546Reputation: 4546Reputation: 4546
If power consumption is a concern you might want to look at low voltage RAM, like Kingston's HyperX LoVo or any other RAM that confirms with the JEDEC specification DDR3L.
 
Old 11-07-2012, 04:21 PM   #48
Ulysses_
Member
 
Registered: Jul 2009
Posts: 756

Original Poster
Rep: Reputation: 43
Does this work immediately or must you set any jumpers on the mobo?

Also, how much does its wattage differ from other DDR3-1600 options? 32GB will probably be installed.

Last edited by Ulysses_; 11-07-2012 at 04:22 PM.
 
Old 11-07-2012, 04:42 PM   #49
TobiSGD
Moderator
 
Registered: Dec 2009
Location: Germany
Distribution: Whatever fits the task best
Posts: 16,511
Blog Entries: 2

Rep: Reputation: 4546Reputation: 4546Reputation: 4546Reputation: 4546Reputation: 4546Reputation: 4546Reputation: 4546Reputation: 4546Reputation: 4546Reputation: 4546Reputation: 4546
It should work immediately. I personally have never used low voltage RAM, so I cant give you data on that. May be this can give you an idea about what we are talking here: http://www.tomshardware.com/reviews/...er,2650-7.html
Keep in mind that this article is from 2010, it may be that power efficiency is better nowadays.
 
Old 11-07-2012, 06:08 PM   #50
Ulysses_
Member
 
Registered: Jul 2009
Posts: 756

Original Poster
Rep: Reputation: 43
If LoVo saves up to 3W at 4GB, what should it save at 32GB?

Or is it the number of sticks that matters, not the GB, as this psu calculator seems to think?

Last edited by Ulysses_; 11-07-2012 at 06:15 PM.
 
Old 11-07-2012, 06:27 PM   #51
TobiSGD
Moderator
 
Registered: Dec 2009
Location: Germany
Distribution: Whatever fits the task best
Posts: 16,511
Blog Entries: 2

Rep: Reputation: 4546Reputation: 4546Reputation: 4546Reputation: 4546Reputation: 4546Reputation: 4546Reputation: 4546Reputation: 4546Reputation: 4546Reputation: 4546Reputation: 4546
I would rather think that it is the number of sticks that matters, not the GB.
Of course, not having RAM will save the most amount of power, so if you in fact only need 16GB for your workload, putting in 32GB will be a greater waste of power as you can get back with using LoVo RAM. Same of course with the SSDs/RevoDrive and any other component in the system. If you don't need the speed of several SSDs or the revodrive then you will get the best powersave effect (and money saving also) if you simply don't use them.
To really get the most out of it you should run a test system to evaluate your workload.
 
Old 11-08-2012, 01:25 AM   #52
cascade9
Senior Member
 
Registered: Mar 2011
Location: Brisneyland
Distribution: Debian, aptosid
Posts: 3,718

Rep: Reputation: 904Reputation: 904Reputation: 904Reputation: 904Reputation: 904Reputation: 904Reputation: 904Reputation: 904
Quote:
Originally Posted by Ulysses_ View Post
Luckily the 125mm tall NH-U9B SE2 is also significantly quieter than the bigger NH-U12P SE. A bit of a paradox here, smaller fans are usually noisier.
Not a paradox at all. Noctua knows what they are doing with fans.

Quote:
Originally Posted by Ulysses_ View Post
Can't I put 2.5" drives in 3.5" slots?
Yes. You need a 2.5'' to 3.5'' adapter for them to fit securely in most cases.

Quote:
Originally Posted by Ulysses_ View Post
So the motherboard's raid was only mentioned to make a higher capacity with mechanical hard disks. Do you have any recommendations for mechanical drives to make 4TB or more? Or a single drive, but I can't find a low-wattage 2.5" hard drive at that capacity.
You cant get 2.5'' HDD bigger than 2.0TB currently.

Quote:
Originally Posted by Ulysses_ View Post
No reply from that shop regarding i7-3770T availability. You said the i7-3770K can be made to run slower in order to lower the wattage. The i7-3770K is a supported cpu for the motherboard, does the motherboard allow under-clocking it? Is that something that can be done with a script while the system is on?
GA-Q77M-D2H seems to have some over/underclocking options. How well they would work is beyond me, Q77 chipsets arent made for over/underclocking.

Quote:
Originally Posted by Ulysses_ View Post
Anyway, the i7-3770K is only slightly faster than the i7-3770T but much more energy hungry. So I took the risk and ordered an i7-3770T from a German shop, fingers crossed.
'Slightly faster' from a passmark forum admin.....and appling what he is saying to your situation could be wrong (and to be truthful 'David (PassMark)' is wrong, or at least misleading as well).

i7-3770K- 2.5GHz, 3.7GHz turbo.
i7-3770T- 3.5GHz, 3.9GHz turbo.

If you only care about single core preformance/max turbo speed, yes, the i7-3770T is only 'slightly faster'.

If you have several cores under load, the i7-3770K is up to 40-50% faster.

You seem to be happy enough to accept advice from a forum as long as it agrees with your ideas.

Quote:
Originally Posted by Ulysses_ View Post
The concern with raid performance is that it is a journey into the unknown, until you buy the hardware or find someone online who has tested the same motherboard and a given set of SSD's, and chances are their SSDs are not high-end and the scores are not as high as the revo's. Whereas the revo with its onboard raid-0 is much more predictable.
Even if you find an online review that doesnt mean its accurate. Like I said earlier, payola is everywhere.

Quote:
Originally Posted by TobiSGD View Post
It should work immediately. I personally have never used low voltage RAM, so I cant give you data on that. May be this can give you an idea about what we are talking here: http://www.tomshardware.com/reviews/...er,2650-7.html
I cant 100% remember right now but IIRC tomshardware uses 'power consumption from the wall' method (i.e. doesnt count the inefficiency of the PSU).

IMO the difference in real system power consumption between the lovo sticks and normal sitcks is less than the tomshardware numbers could lead you to believe.

Xbitlabs uses an 'after the PSU power consumption' method which is more accurate IMO-

http://www.xbitlabs.com/articles/mem...vo-hyperx.html

Still, provided they run with ivybidge (havent checked, I'd guess they should) and arent to expensive its not a bad idea..even saying 2-6watts is a bonus.

Last edited by cascade9; 11-08-2012 at 07:20 AM. Reason: typo
 
Old 11-08-2012, 06:59 AM   #53
Ulysses_
Member
 
Registered: Jul 2009
Posts: 756

Original Poster
Rep: Reputation: 43
Risking with a cpu that needs underclocking and without knowing if the mobo's underclocking was sufficient, this seemed too much of a risk.

And by the way, the single-threaded test is on a separate page here, where the i7-3770K is 5% faster. The multi-threaded test is here, where the i7-3770K is only 10% faster. The passmark tests are multi-threaded according to this.

Regarding Kingston LoVo, page 3 of the test site says: "Many mainboards, such as Gigabyte ones, for instance, can’t set such low voltages. At least not without the special BIOS modifications." How do we find out if the BIOS of the motherboard has settings for LoVo memory?

Last edited by Ulysses_; 11-08-2012 at 04:14 PM.
 
Old 11-08-2012, 11:14 AM   #54
Ulysses_
Member
 
Registered: Jul 2009
Posts: 756

Original Poster
Rep: Reputation: 43
Found the motherboard's manual here, it says something about memory voltage. Here's the critical part from pages 29 to 30:

http://www.linuxquestions.org/questi...2&d=1352416225

Is DDR3L definitely supported?

Last edited by Ulysses_; 11-08-2012 at 05:32 PM.
 
Old 11-08-2012, 03:48 PM   #55
jefro
LQ Guru
 
Registered: Mar 2008
Posts: 13,174

Rep: Reputation: 1667Reputation: 1667Reputation: 1667Reputation: 1667Reputation: 1667Reputation: 1667Reputation: 1667Reputation: 1667Reputation: 1667Reputation: 1667Reputation: 1667
The core i7 3770 has more virtual machine support.
 
Old 11-08-2012, 05:15 PM   #56
Ulysses_
Member
 
Registered: Jul 2009
Posts: 756

Original Poster
Rep: Reputation: 43
Kingston DDR3L is available in CL9 and CL10 latencies, where the former is 50% more expensive. Is it worth it?
Attached Thumbnails
Click image for larger version

Name:	Image4.jpg
Views:	53
Size:	151.9 KB
ID:	11182  

Last edited by Ulysses_; 11-08-2012 at 05:30 PM.
 
Old 11-09-2012, 01:26 AM   #57
cascade9
Senior Member
 
Registered: Mar 2011
Location: Brisneyland
Distribution: Debian, aptosid
Posts: 3,718

Rep: Reputation: 904Reputation: 904Reputation: 904Reputation: 904Reputation: 904Reputation: 904Reputation: 904Reputation: 904
Quote:
Originally Posted by Ulysses_ View Post
Risking with a cpu that needs underclocking and without knowing if the mobo's underclocking was sufficient, this seemed too much of a risk.
You dont even know if you 'need' a i7-3770T or the lower TDP/max power consumption benfits.

Quote:
Originally Posted by Ulysses_ View Post
And by the way, the single-threaded test is on a separate page here, where the i7-3770K is 5% faster. The multi-threaded test is here, where the i7-3770K is only 10% faster. The passmark tests are multi-threaded according to this.
Do you have any idea what passmark is and how it works?

It is in no way a benchmarking compariosn system. You cant use it as a benchmark.

Pasmark is d/l and run on whatever hrdware the user has. For any single CPU, you can have various different RAM sizes, speeds, CAS ratings, chipsets, HDDs/SSDs, video cards (or not) etc..

Given enough samples, it gives at best a rough guide.

i7-3770T- 3 samples. Statistically insignificant.

I'm not suprised taht there are so few i7-3770T passmark results, its an very uncommon and hard to find CPU.

Even if you believe that 3 samples is statistically significant, and the variance in testing systems doesnt matter, that doesnt mean that passmark is doing a good job on measuring multicore performance.

While I cant find any decent test on the i7-3770T, more than 10% difference between the the i7-3770T and i7-3770K would be in line with results I've seen for other lower power Intel CPUs. I'd gues than on average it would be 10-20% and I believe than in the right circumstances it would be as high as 40-50%.

If you want to believe legionhardware tests becuase they fit with your conceptions, or use passmark to show performance differences without understanding how passmark works, that is up to you.
 
Old 11-09-2012, 05:34 AM   #58
Ulysses_
Member
 
Registered: Jul 2009
Posts: 756

Original Poster
Rep: Reputation: 43
One has to start somewhere and the only available place for me was google with the words "benchmark" or "score", for any recommended cpu. There is no pre-conception at all.
 
Old 11-09-2012, 07:34 AM   #59
Ulysses_
Member
 
Registered: Jul 2009
Posts: 756

Original Poster
Rep: Reputation: 43
Quote:
Originally Posted by jefro View Post
The core i7 3770 has more virtual machine support.
Is there something about virtualization that is not said at intel's pages about this cpu? Intel's comparison tool shows these results:

http://ark.intel.com/compare/65523,65719,65525

From this comparison it seems that the 3770T has all the features of the 3770, except one: "embedded options". It's probably an underclocked 3770. Whereas the 3770K is missing some virtualization features that the other two have.

Last edited by Ulysses_; 11-09-2012 at 08:32 AM.
 
Old 11-10-2012, 06:16 AM   #60
Ulysses_
Member
 
Registered: Jul 2009
Posts: 756

Original Poster
Rep: Reputation: 43
In the future, if even higher storage performance is required, what sort of gain do you expect from software raid-0 with a second revo drive?

Assuming it will work on the gpu's pcie x16 slot, is that correct?
 
  


Reply


Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is Off
HTML code is Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
PCIe re-enumation in linux driver question (pcie hotplug doesn't work) blavo Programming 5 08-02-2012 02:12 PM
Intel SSD Series 30/311 PCIe Questions ThomasLMcLean Linux - Hardware 0 07-04-2011 11:16 AM
home storage server question: low power consumption features... Yaniv-Fer Linux - General 7 03-01-2009 02:05 AM
Building or buying a HDD-less server for low power consumption bence8810 Linux - Hardware 2 12-12-2007 06:51 AM
Low power consumption file server? Lordandmaker General 7 04-27-2006 09:34 PM


All times are GMT -5. The time now is 09:57 PM.

Main Menu
Advertisement
My LQ
Write for LQ
LinuxQuestions.org is looking for people interested in writing Editorials, Articles, Reviews, and more. If you'd like to contribute content, let us know.
Main Menu
Syndicate
RSS1  Latest Threads
RSS1  LQ News
Twitter: @linuxquestions
Facebook: linuxquestions Google+: linuxquestions
Open Source Consulting | Domain Registration