LinuxQuestions.org
Download your favorite Linux distribution at LQ ISO.
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 03-24-2014, 06:31 AM   #1
Ulysses_
Senior Member
 
Registered: Jul 2009
Posts: 1,303

Rep: Reputation: 57
At the same PCmark score, is an i5 better than an i7?


In benchmark comparisons it often turns out that an intel i7 cpu has about the same score as an i5. Does this mean the i5 is better than the i7 for most people because most people are running single-threaded applications anyway?

For the purposes of running 3-4 VM's at the same time, each running a browser looking at a heavy site or a video viewer, which is better?

Last edited by Ulysses_; 03-24-2014 at 06:35 AM.
 
Old 03-24-2014, 06:40 AM   #2
Germany_chris
Senior Member
 
Registered: Jun 2011
Location: NOVA
Distribution: Debian 12
Posts: 1,069

Rep: Reputation: 497Reputation: 497Reputation: 497Reputation: 497Reputation: 497
i7..

You need cores either real or fake for multiple VM's you also need virtualization support which is not available in all i5's. I personally went with an E3 Xeon (E3-1265L) they are price competitive and have all the features. If you're looking at K SKU's none have virtualization support. If you're doing a new build have a look at AMD
 
Old 03-24-2014, 08:04 AM   #3
Ulysses_
Senior Member
 
Registered: Jul 2009
Posts: 1,303

Original Poster
Rep: Reputation: 57
If only 3-4 VM's are run at a time and an i5 is found with virtualization support? Won't each VM run faster than on an i7 with the same pcmark score?
 
Old 03-24-2014, 11:10 AM   #4
Germany_chris
Senior Member
 
Registered: Jun 2011
Location: NOVA
Distribution: Debian 12
Posts: 1,069

Rep: Reputation: 497Reputation: 497Reputation: 497Reputation: 497Reputation: 497
No the i5 is 4 cores 4 threads and the i7 is 4 cores and 8 threads which the computer see's as 8 cores. When you set you VM's up you can give each more cores because you have more
 
Old 03-24-2014, 12:01 PM   #5
TobiSGD
Moderator
 
Registered: Dec 2009
Location: Germany
Distribution: Whatever fits the task best
Posts: 17,148
Blog Entries: 2

Rep: Reputation: 4886Reputation: 4886Reputation: 4886Reputation: 4886Reputation: 4886Reputation: 4886Reputation: 4886Reputation: 4886Reputation: 4886Reputation: 4886Reputation: 4886
Quote:
Originally Posted by Ulysses_ View Post
If only 3-4 VM's are run at a time and an i5 is found with virtualization support? Won't each VM run faster than on an i7 with the same pcmark score?
If you run 4 VMs an a four core machine those 4 VMs have to compete more with the host OS about CPU time as when you run the same workload on a CPU with 4 cores and hyperthreading. Also, if both CPUs have about the same PCMark score (which is a bad benchmark per se, IMHO), why should the i5 suddenly be faster?
 
Old 03-24-2014, 12:05 PM   #6
Germany_chris
Senior Member
 
Registered: Jun 2011
Location: NOVA
Distribution: Debian 12
Posts: 1,069

Rep: Reputation: 497Reputation: 497Reputation: 497Reputation: 497Reputation: 497
By no metric is the i5 faster than the i7 I still thing the OP should get a Xeon.
 
Old 03-24-2014, 04:45 PM   #7
Ulysses_
Senior Member
 
Registered: Jul 2009
Posts: 1,303

Original Poster
Rep: Reputation: 57
Is it accurate that PCmark puts a high load on all cores and tells you the total operations per second by adding the operations of core 1 plus the ops of core 2 plus core 3 etc? So if the sum of an i7 is the same as the sum of an i5, individual cores would be slower in the cpu with the most cores, ie the i7?

This matters because most people are only stressing one core as most loads are single-threaded loads. The process viewer in windows confirms this, it never hits the max in an i7-3770T with a single application but hardly reaches 10% or so. Unless you encode video with a multi-threaded codec, this was the only way I could get it to 100%.

Was going to allocate one core per VM and leave a core for the host. Tried giving the VM's plenty of cores on an i7-3770T (it allowed up to 12 or 16 if I remember well) but the VMs slowed down. What should be the optimum number of cores per VM if only 3 VM's are required?

By the way, cpu's are being compared for laptops as I want to buy one, not sure a xeon is available in a laptop.

Last edited by Ulysses_; 03-24-2014 at 05:00 PM.
 
Old 03-24-2014, 05:00 PM   #8
TobiSGD
Moderator
 
Registered: Dec 2009
Location: Germany
Distribution: Whatever fits the task best
Posts: 17,148
Blog Entries: 2

Rep: Reputation: 4886Reputation: 4886Reputation: 4886Reputation: 4886Reputation: 4886Reputation: 4886Reputation: 4886Reputation: 4886Reputation: 4886Reputation: 4886Reputation: 4886
Quote:
Originally Posted by Ulysses_ View Post
Is it accurate that PCmark puts a high load on all cores and threads and tells you the total operations per second by adding the operations of core 1 plus the ops of core 2 plus core 3 etc? So if the sum of an i7 is the same as the sum of an i5, individual cores would be slower in the cpu with the more cores, ie the i7?
No, PCMark runs several workloads of typical Windows software (mostly Adobe and Microsoft) together with a few synthetic benchmarks, give any of them a score and uses those scores to calculate a general score.
 
Old 03-24-2014, 05:02 PM   #9
Ulysses_
Senior Member
 
Registered: Jul 2009
Posts: 1,303

Original Poster
Rep: Reputation: 57
So no summation is going on?
 
Old 03-24-2014, 08:14 PM   #10
metaschima
Senior Member
 
Registered: Dec 2013
Distribution: Slackware
Posts: 1,982

Rep: Reputation: 492Reputation: 492Reputation: 492Reputation: 492Reputation: 492
i5 is better value than i7, but not better performance. I have an i5 because I don't think the extra performance and hyper-threading is worth $100 more. In most applications (not highly threaded) there won't be much of a difference.
 
Old 03-24-2014, 08:19 PM   #11
jefro
Moderator
 
Registered: Mar 2008
Posts: 21,939

Rep: Reputation: 3619Reputation: 3619Reputation: 3619Reputation: 3619Reputation: 3619Reputation: 3619Reputation: 3619Reputation: 3619Reputation: 3619Reputation: 3619Reputation: 3619
Pretty much, you get what you pay for. If the core i7 costs more, it is better.

Your use and apps would decide if the extra money is worth it. Some features like VM support may not be easily found in common tests. If you are just watching Hulu then I'd think the core i3 would be fine.
 
Old 03-24-2014, 09:06 PM   #12
TobiSGD
Moderator
 
Registered: Dec 2009
Location: Germany
Distribution: Whatever fits the task best
Posts: 17,148
Blog Entries: 2

Rep: Reputation: 4886Reputation: 4886Reputation: 4886Reputation: 4886Reputation: 4886Reputation: 4886Reputation: 4886Reputation: 4886Reputation: 4886Reputation: 4886Reputation: 4886
Quote:
Originally Posted by Ulysses_ View Post
So no summation is going on?
No, they use some arbitrary rules to determine the end result from the results of the single tests. On the Wikipedia page for PCMark you can find a brief description of those test, but I guess the rules to determine the end result are secret.
 
Old 03-25-2014, 05:06 AM   #13
Ulysses_
Senior Member
 
Registered: Jul 2009
Posts: 1,303

Original Poster
Rep: Reputation: 57
What's a better benchmark then, that is relevant when you want to run 3 or 4 VM's, and one that has many cpu's tested and published online for comparison?
 
Old 03-25-2014, 05:10 AM   #14
Ulysses_
Senior Member
 
Registered: Jul 2009
Posts: 1,303

Original Poster
Rep: Reputation: 57
PS The VM's are running browsers visiting client-heavy sites and playing videos.
 
Old 03-25-2014, 05:50 AM   #15
TobiSGD
Moderator
 
Registered: Dec 2009
Location: Germany
Distribution: Whatever fits the task best
Posts: 17,148
Blog Entries: 2

Rep: Reputation: 4886Reputation: 4886Reputation: 4886Reputation: 4886Reputation: 4886Reputation: 4886Reputation: 4886Reputation: 4886Reputation: 4886Reputation: 4886Reputation: 4886
I don't know of benchmarks using VMs, but if total CPU power is your concern benchmarks like the multithreading part of CineBench (rendering 3D scenes) or wprime (calculating large prime numbers) come to my mind. If they really are relevant for you is something I can't say, the only reliable benchmark would be to test your workload on both CPUs.

From a different perspective, assuming that you are paid to do the research for building that machine, the costs of your research time now should have excelled the extra costs for the i7.
 
  


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
[SOLVED] X-Score Maat84 Programming 1 02-19-2010 06:07 AM
Spamassassin score alitabas Linux - Software 1 10-15-2009 02:17 PM
Newbie lookin' to score boygenuis LinuxQuestions.org Member Intro 3 10-27-2007 10:32 AM
Credit score KenMarshall General 1 04-08-2007 12:39 PM
What license should SCORE use? flamesrock Linux - Software 2 02-08-2004 09:29 PM

LinuxQuestions.org > Forums > Linux Forums > Linux - Hardware

All times are GMT -5. The time now is 07:02 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
Open Source Consulting | Domain Registration