Home | Blog | Software | Reviews and Features | Forum | Help | Donate | About us
topbanner_forum
  *

avatar image

Welcome, Guest. Please login or register.
Did you miss your activation email?

Login with username, password and session length
  • September 02, 2015, 09:32:53 PM
  • Proudly celebrating 10 years online.
  • Donate now to become a lifetime supporting member of the site and get a non-expiring license key for all of our programs.
  • donate

Author Topic: Test: Does latency reduction via RAM upgrade lift software performance?  (Read 2406 times)

IainB

  • Supporting Member
  • Joined in 2008
  • **
  • Posts: 5,432
  • Slartibartfarst
    • View Profile
    • Donate to Member
I wanted to know whether latency reduction via RAM upgrade would lift software performance.
So I did a live test on two laptops.
The conclusion is that RAM upgrades can make a significant improvement to latency (reducing it) and are well worth doing.

A few years back, I had used RAM upgrades to a couple of Toshiba (Windows XP) laptops (now defunct), taking them from the standard 512Mb or 1Gb up to 3 or 4Gb. The difference in "feel" was quite impressive. It felt much faster on most large applications, though I had no easy way of measuring this.
A couple of months back, I helped out some neighbours whose Asus (Windows Vista) laptop had been hijacked by a trojan virus. After I had restored it to normal, I thought it was a bit slow, and swapped-in 3Gb of DDR2 RAM from a defunct Toshiba. It seemed much better.
In fact, the transformation was impressive, and the owners were really pleased with the result.

I had purchased my HP ENVY 14 laptop over a year ago, and my daughter's (Lily's) DELL Inspiron laptop about 8 months ago. Looking at laptop DDR3 4Gb RAM card prices, I noticed they were affordable at about NZ$50 each. So I decided to upgrade the two laptops and measure the resulting performance changes.

Currently, laptops seem to come with only two slots for DDR3 RAM.
The HP had 1x4Gb card only, with the other slot empty.
The DELL had 2x2Gb cards - one in each slot.

HP upgrade:
  • I put one 4Gb card into the empty slot, giving 8Gb in total, and then ran the standard Windows Experience Index performance assessment tool.

DELL upgrade:
  • I took out one of the DELL's (2Gb) cards, and replaced it with a 4Gb card, giving 6Gb in total, and then ran the standard Windows Experience Index performance assessment tool.
  • Then I took out the remaining one of the DELL's 2Gb cards and replaced it with a 4Gb card, giving 8Gb in total, and then ran the standard WEI performance assessment tool again.

From the user's perspective on the HP, I noticed significantly reduced latency in almost any program being run, and especially in MS OneNote, Firefox, Chrome/Chromium, InfoSelect, and some xplorer² functionality.

From the user's perspective on the DELL, Lily noticed delays being significantly reduced in loading/closing SIMS3 games and in other functionality in SIMS3 and SIMS Medieval, and in online games and Firefox. She said "It was all much faster".

The Windows Experience Index performance assessment results help to explain this, as below:
DDR3 RAM upgrade performance table.jpg
« Last Edit: April 13, 2012, 11:25:17 PM by IainB, Reason: Minor corrections. »

techidave

  • Supporting Member
  • Joined in 2007
  • **
  • Posts: 992
    • View Profile
    • Pawnee Heights Tigers
    • Donate to Member
Is the Windows Experience Index an accurate tool to measure with?  Its interesting to note that the base score didn't improve much if any.  So I wonder, what kind of upgrade can one do to increase the base score?

IainB

  • Supporting Member
  • Joined in 2008
  • **
  • Posts: 5,432
  • Slartibartfarst
    • View Profile
    • Donate to Member
Is the Windows Experience Index an accurate tool to measure with?  Its interesting to note that the base score didn't improve much if any.  So I wonder, what kind of upgrade can one do to increase the base score?
I don't know whether the WEI is a particularly accurate tool, or if it is the best or most accurate tool to measure with. I am sure there could be other benchmark tools that you could probably use. I just used the one provided by Microsoft because it was an easy no-fuss standard tool that provided a reasonably consistent basis of comparison - which I guess was one of the purposes it was designed for.

The Base Score is not an average of all the subscores. It is just the lowest subscore from the set - which makes sense, because the system component with the lowest subscore would probably be a potential system bottleneck in some way.
In neither case - the HP or the DELL - was the measure that gave the base score significantly affected by the RAM upgrade. Hence, no significant change to base score. But what the RAM upgrade did do was significantly and perceptibly reduce the latency in both cases - and the measure of why is given by the greatly increased subscore for memory operations.

My objective was to establish whether the latency reduction via RAM upgrade would lift software performance.
The conclusion was that RAM upgrades can make a significant improvement to latency (reducing it) and are therefore probably well worth doing.

I was thus not focussed per se on raising the lowest performance subscore that currently made for the Base Score.
One thought I had from the test was that it might be worth tinkering with the system settings for the Aero and/or graphics functionality subscores on the DELL laptop. That might be something that could be lifted with a bit of tuning. At the moment I am unsure as to why it is relatively lower on the DELL compared to the same subscores on the HP laptop.

x16wda

  • Supporting Member
  • Joined in 2007
  • **
  • Posts: 591
  • what am I doing in this handbasket?
    • View Profile
    • Donate to Member
I have seen most often the lowest-performing subsystem is graphics, however with pressure on hard drives due to the flood in Thailand most manufacturers are using fairly slow drives.  It's easier to get a faster hard drive on a laptop (i'm saving up for a hybrid for mine), somewhat harder to upgrade the video...
vi vi vi - editor of the beast

IainB

  • Supporting Member
  • Joined in 2008
  • **
  • Posts: 5,432
  • Slartibartfarst
    • View Profile
    • Donate to Member
After using these laptops for some time with the RAM upgrades, the experience continues to be that "everything seems to run faster".
In an effort to improve the screen display quality I have also now turned on (previously turned off most) all the settings for max graphics quality - on both laptops - which, theoretically should place extra load on the CPUs. However, the user experience is that display quality/resolution has improved and there has been no perceptible waiting/latency from any processes.

I have also switched off any GPU tweaks (both laptops have ATI Radeon GPUs), as I read on 3Dguru that the GPUs generally only achieve performance benefits by lossy methods. In our case, the user experience is that display quality/resolution seems to have slightly improved and there has been no perceptible waiting/latency from any processes. If you were into RPG (utilising lots of graphics load on the GPU), then I suppose that you might notice a difference.

CONCLUSION so far: This real-world test overall seems to have provided some pretty good empiric evidence that the best and cheapest major performance improvement you can probably give any PC or laptop is arguably likely be a RAM upgrade. I suspect that it could seem to give new life to older PCs and help to defer their replacement date.

IainB

  • Supporting Member
  • Joined in 2008
  • **
  • Posts: 5,432
  • Slartibartfarst
    • View Profile
    • Donate to Member

Carol Haynes

  • Waffles for England (patent pending)
  • Global Moderator
  • Joined in 2005
  • *****
  • Posts: 7,969
    • View Profile
    • Dales Computer Services
    • Donate to Member
The base score isn't going to change in the WE Tool. You cannot get above 5.9 unless you have an SSD installed. How is upgrading RAM suposed to affect Graphics performance - basically what it is saying is that the GPU is underpowered (true on most domestic laptops).

The biggest change by increasing RAM is to reduce pagefile swapping which is a huge burden on heavily used machines. There is an argument to turn of the pagefile altogether if you have 8Gb of RAM or more - not sure if this is such a good idea or not (never tried it) but you could try with a minimal pagefile size (to allow for minidumping on error).