Forums » Technology » Hardware & Software » 512MB, 1024MB, and 2048MB of RAM

Page 1 of 5 1 2 3 4 5 >
Topic Options
Rate This Topic
Hop to:
#124711 - 10/20/06 08:43 PM 512MB, 1024MB, and 2048MB of RAM
Joe Offline
SimHQ Technology Editor
SimHQ Lifer

Registered: 04/05/02
Posts: 16196
Loc: Somerville, NJ
SimHQ originally intended to publish the benchmark results below as a technology article. However, the results aren't really that surprising and it's probably not worth the effort it would take to format such a piece. Still, the numbers are interesting and useful, so they have been posted here as a "mini-article". Please use this thread to comment.

Test System
CPU: Opteron 148 @ 2200 MHz
Motherboard: Asrock 939Dual-SATAII
GPU: 256MB PCI-E 7900GT using Forceware 91.33
Memory: 1x, 2x, or 4x 512MB 400MHz DDR (Corsair XMS3200 v1.2) (1x512 is a single channel configuration while 2x and 4x512 are dual channel configurations)

Sims tested
  • IL-2 Forgotten Battles + Pacific Fighters + Aces Expansion Pack 4.04
  • Lock On + Flaming Cliffs 1.12a
  • Falcon 4: Allied Force 1.07
  • Wings Over Vietnam 07.10.06
  • rFactor 1150


Test settings
All sims were tested at 1280x960 (exception: rFactor tested at 1024x768) with 4xAA and 8xAF forced on for all tests. The goal was to place a reailstic but not excessive load on the video card, in order to recreate realistic game settings. RAM was set to two different speed/latency combinations; the first was 266MHz with 3-6-6-15-2T timings (representative of older memory), and the second was 400 MHz with 2.5-3-3-11-2T timings (representing current average DDR settings).

Tests
The following were evaluted for each of the five sims tested:
  • Sim load time from the desktop to the sim's user interface
  • Mission or track load time from the user interface to the game's 3D environment
  • Average/minimum/and maximum framerates during a session, or instantaneous framerate in-game



RESULTS: FB+PF+AEP 4.04 load time from desktop
1x512MB (512MB total), 266 MHz, 3-6-6-15-2T: 46 seconds
1x512MB (512MB total), 400 MHz, 2.5-3-3-11-2T: 36 seconds
2x512MB (1024MB total), 400 MHz, 2.5-3-3-11-2T: 25 seconds
4x512MB (2048MB total), 400 MHz, 2.5-3-3-11-2T: 24 seconds

RESULTS: FB+PF+AEP 4.04 Black Death track load time
1x512MB (512MB total), 266 MHz, 3-6-6-15-2T: 26 seconds
1x512MB (512MB total), 400 MHz, 2.5-3-3-11-2T: 26 seconds
2x512MB (1024MB total), 400 MHz, 2.5-3-3-11-2T: 11 seconds
4x512MB (2048MB total), 400 MHz, 2.5-3-3-11-2T: 11 seconds

RESULTS: FB+PF+AEP 4.04 Black Death track average/minimum/maximum framerates
1x512MB (512MB total), 266 MHz, 3-6-6-15-2T: 23/3/40
1x512MB (512MB total), 400 MHz, 2.5-3-3-11-2T: 29/11/48
2x512MB (1024MB total), 400 MHz, 2.5-3-3-11-2T: 32/12/51
4x512MB (2048MB total), 400 MHz, 2.5-3-3-11-2T: 37/13/51

The IL-2 series sees a lot of improvement from a move from low- to high-speed RAM; there is a significant decrease in game loading time and a significant increase in minimum framerate. A jump can also be seen from 512MB to 1024MB; game load time drops again and this time track load time does as well. Framerates don't change much, however. Moving from 1024MB to 2048MB results in very little change anywhere.


RESULTS: Lock On + Flaming Cliffs 1.12a load time from desktop
1x512MB (512MB total), 266 MHz, 3-6-6-15-2T: 29 seconds
1x512MB (512MB total), 400 MHz, 2.5-3-3-11-2T: 29 seconds
2x512MB (1024MB total), 400 MHz, 2.5-3-3-11-2T: 25 seconds
4x512MB (2048MB total), 400 MHz, 2.5-3-3-11-2T: 24 seconds

RESULTS: Lock On + Flaming Cliffs 1.12a MiG-29 demo track load time
1x512MB (512MB total), 266 MHz, 3-6-6-15-2T: 76 seconds
1x512MB (512MB total), 400 MHz, 2.5-3-3-11-2T: 53 seconds
2x512MB (1024MB total), 400 MHz, 2.5-3-3-11-2T: 21 seconds
4x512MB (2048MB total), 400 MHz, 2.5-3-3-11-2T: 21 seconds

RESULTS: Lock On + Flaming Cliffs 1.12a MiG-29 demo average/minimum/maximum framerates
1x512MB (512MB total), 266 MHz, 3-6-6-15-2T: 23/3/40
1x512MB (512MB total), 400 MHz, 2.5-3-3-11-2T: 44/17/86
2x512MB (1024MB total), 400 MHz, 2.5-3-3-11-2T: 45/24/87
4x512MB (2048MB total), 400 MHz, 2.5-3-3-11-2T: 47/25/109

While Lock On's game load times remain relatively unaffected by memory parameters, massive improvements can be seen in mission load times when increasing the speed of 512MB memory and moving to 1024MB. The same can be said about framerates; minimum framerates increase from 3fps to 17 fps and then to 24 fps. Neglible improvements are seen from a move to 2048MB. Lock On is an extreme example of a game that has a 1024MB "sweet spot"; a 53-second mission load time compared to a 21-second load time is the difference between twiddling your thumbs/rolling your eyes and a reasonable wait.


RESULTS: Falcon 4:Allied Force 1.07 load time from desktop
1x512MB (512MB total), 266 MHz, 3-6-6-15-2T: 32 seconds
1x512MB (512MB total), 400 MHz, 2.5-3-3-11-2T: 26 seconds
2x512MB (1024MB total), 400 MHz, 2.5-3-3-11-2T: 24 seconds
4x512MB (2048MB total), 400 MHz, 2.5-3-3-11-2T: 24 seconds

RESULTS: Falcon 4:Allied Force 1.07 Day 1 Balkans H-hour +2 SEAD escort mission load time
1x512MB (512MB total), 266 MHz, 3-6-6-15-2T: 35 seconds
1x512MB (512MB total), 400 MHz, 2.5-3-3-11-2T: 40 seconds
2x512MB (1024MB total), 400 MHz, 2.5-3-3-11-2T: 25 seconds
4x512MB (2048MB total), 400 MHz, 2.5-3-3-11-2T: 25 seconds

RESULTS: Falcon 4:Allied Force 1.07 Day 1 Balkans H-hour +2 SEAD escort mission instantaneous framerate in cockpit on runway
1x512MB (512MB total), 266 MHz, 3-6-6-15-2T: 20
1x512MB (512MB total), 400 MHz, 2.5-3-3-11-2T: 28
2x512MB (1024MB total), 400 MHz, 2.5-3-3-11-2T: 28
4x512MB (2048MB total), 400 MHz, 2.5-3-3-11-2T: 29

AF shows a slight improvement moving from slow to fast memory, but beyond that not much effect can be seen. The CPU likely is the limiting hardware here.


RESULTS: Wings Over Vietnam 07.10.06 load time from desktop
1x512MB (512MB total), 266 MHz, 3-6-6-15-2T: 17 seconds
1x512MB (512MB total), 400 MHz, 2.5-3-3-11-2T: 17 seconds
2x512MB (1024MB total), 400 MHz, 2.5-3-3-11-2T: 9 seconds
4x512MB (2048MB total), 400 MHz, 2.5-3-3-11-2T: 8 seconds

RESULTS: Wings Over Vietnam 07.10.06 Operation Bolo single mission load time
1x512MB (512MB total), 266 MHz, 3-6-6-15-2T: 17 seconds
1x512MB (512MB total), 400 MHz, 2.5-3-3-11-2T: 19 seconds
2x512MB (1024MB total), 400 MHz, 2.5-3-3-11-2T: 17 seconds
4x512MB (2048MB total), 400 MHz, 2.5-3-3-11-2T: 18 seconds

RESULTS: Wings Over Vietnam 07.10.06 Operation Bolo single mission instantaneous framerate in cockpit in flight
1x512MB (512MB total), 266 MHz, 3-6-6-15-2T: 55
1x512MB (512MB total), 400 MHz, 2.5-3-3-11-2T: 67
2x512MB (1024MB total), 400 MHz, 2.5-3-3-11-2T: 72
4x512MB (2048MB total), 400 MHz, 2.5-3-3-11-2T: 73

A large decrease in load time can be seen moving from 512MB to 1024MB of RAM, and framerates jump when increasing RAM speed as well as moving from 512MB to 1024MB. However, it's doubtful you'll be able to tell the difference between 55fps and 67fps. WOV also had the shortest load time of any sim so an improvement here is not all that important. There is no significant improvement moving from 1024MB to 2048MB.


RESULTS: rFactor 1150 load time from desktop
1x512MB (512MB total), 266 MHz, 3-6-6-15-2T: 17 seconds
1x512MB (512MB total), 400 MHz, 2.5-3-3-11-2T: 17 seconds
2x512MB (1024MB total), 400 MHz, 2.5-3-3-11-2T: 9 seconds
4x512MB (2048MB total), 400 MHz, 2.5-3-3-11-2T: 8 seconds

RESULTS: rFactor 1150 BMW Sauber practice @ Silverstone (Northamptonshire) track load time
1x512MB (512MB total), 266 MHz, 3-6-6-15-2T: 17 seconds
1x512MB (512MB total), 400 MHz, 2.5-3-3-11-2T: 19 seconds
2x512MB (1024MB total), 400 MHz, 2.5-3-3-11-2T: 17 seconds
4x512MB (2048MB total), 400 MHz, 2.5-3-3-11-2T: 18 seconds

RESULTS: rFactor 1150 replay of Lupos @ Lienz Alstadt average/minimum/maximum framerates
1x512MB (512MB total), 266 MHz, 3-6-6-15-2T: 27/14/43
1x512MB (512MB total), 400 MHz, 2.5-3-3-11-2T: 27/14/43
2x512MB (1024MB total), 400 MHz, 2.5-3-3-11-2T: 27/14/43
4x512MB (2048MB total), 400 MHz, 2.5-3-3-11-2T: 27/14/43

Out of the five sims tested, rFactor responds the least to memory changes. Track load times and framerates were unchanged for all memory configurations, and only the game load time improved when moving from 512MB to 1024MB. Just like WOV, this is already a short load time.


CONCLUSION
Significant improvements can be seen in sims when moving from 512MB to 1024MB of RAM. Very little to no improvement is typically seen when moving from 1024MB to 2048MB of RAM. Although there are no doubt specific situations where 2048MB of RAM is helpful, at this time it should not be considered a "must-have."

Top Bookmark and Share
#124712 - 10/20/06 09:56 PM Re: RAM Benchmarks: the effect of 512MB, 1024MB, and 2048MB of memory
Hawkw1nd Offline
SimHQ Member

Registered: 06/17/02
Posts: 3170
Loc: Sydney, AUSTRALIA
Very well done, old chap. A most interesting analysis !!!
_________________________
Charter member of the FanATIcal nVidiots Society of Uranus

Top Bookmark and Share
#124713 - 10/21/06 01:53 AM Re: RAM Benchmarks: the effect of 512MB, 1024MB, and 2048MB of memory
C3PO Offline
SimHQ Member

Registered: 10/29/02
Posts: 663
Very interesting tests!
_________________________
Intel E6600 Core 2 Duo @ 3GHz, 2GB DDR2 633MHz RAM, XFX 8800 GTS 320MB video card -- great for GTR2 and MoTeC i2 Pro analysis software
My Brno telemetry: http://img156.imageshack.us/img156/4245/telemetryv2xq2.jpg

Top Bookmark and Share
#124714 - 10/21/06 05:37 AM Re: RAM Benchmarks: the effect of 512MB, 1024MB, and 2048MB of memory
- Ice Offline
SimHQ Member

Registered: 06/23/05
Posts: 1038
Loc: Philippines
I'm not too computer-savvy, so tell me, if 1024MB of RAM is the "sweet spot" for games nowadays, what happens to the rest of the RAM for guys with 2-3Gig of RAM in their system?
_________________________
- Ice

Top Bookmark and Share
#124715 - 10/21/06 06:26 AM Re: RAM Benchmarks: the effect of 512MB, 1024MB, and 2048MB of memory
Dragon_Tail-115 Offline
SimHQ Member

Registered: 06/08/03
Posts: 291
Loc: israel
Usefull data .
Thanks man.
_________________________
AMD XP 2600+ Gigabyte ga7vax 1Gb memory Radeon 9800 xt 256mb WinXP PRO SP2 TopGun AfterBurner II stick+Throtle
TrackIR.

Top Bookmark and Share
#124716 - 10/21/06 08:23 AM Re: RAM Benchmarks: the effect of 512MB, 1024MB, and 2048MB of memory
Gatticus Offline
SimHQ Member

Registered: 08/06/05
Posts: 2894
Loc: Canada
Quote:
Originally posted by ecienavxela:
I'm not too computer-savvy, so tell me, if 1024MB of RAM is the "sweet spot" for games nowadays, what happens to the rest of the RAM for guys with 2-3Gig of RAM in their system?
It never gets used. But 2gb is useful in some games like Oblivion and BF2.
_________________________
"Electronic music should have freed the masses from the limitations of fretboards and fixed-pitch instruments. But apparently some missed the memo..." --thegibsonian

Top Bookmark and Share
#124717 - 10/21/06 09:50 AM Re: RAM Benchmarks: the effect of 512MB, 1024MB, and 2048MB of memory
Nixer Online   smile
SimHQ Senior Member

Registered: 03/06/01
Posts: 5802
Loc: Lousyanna
Thanks Joe, very interesting and informative.
_________________________
I wish........................

Top Bookmark and Share
#124718 - 10/21/06 11:07 AM Re: RAM Benchmarks: the effect of 512MB, 1024MB, and 2048MB of memory
TheJedi Offline
SimHQ Member

Registered: 06/17/01
Posts: 151
I think such a test run would be very interesting with FSX.

Top Bookmark and Share
#124719 - 10/21/06 11:45 AM Re: RAM Benchmarks: the effect of 512MB, 1024MB, and 2048MB of memory
WalterNowi Offline
SimHQ Member

Registered: 03/08/04
Posts: 370
Loc: San Gabriel, CA
Joe,
Interesting results. I am wondering what the results would look like for:

2 x 512 MB @1T versus
2 x 1024 MB @1T
_________________________
Intel Q6600 B3
Abit IP35 Pro
2 x 2 Gb Corsair XMS2 PC2 6400
EVGA GTX 260 Core 216 Superclocked Edition
Silverstone DA650 650W
CH Products FS/PT/PP/TQ
Track IR3 Pro w/ Vector

Top Bookmark and Share
#124720 - 10/21/06 01:27 PM Re: RAM Benchmarks: the effect of 512MB, 1024MB, and 2048MB of memory
Speedo Offline
SimHQ Member

Registered: 02/15/01
Posts: 4551
Loc: NC USA
It's a pretty good summary of the current flight sim situation, but I'd like to see similar testing done with FSX.
_________________________
Athlon64x2 4400+
Asus A8N-SLI Premium
2Gb Corsair XMS Pro @ 3-3-3-8
MSI GF8800GT 512MB
Creative X-Fi
2x Samsung SATA2 80Gb striped
Seasonic 600W PSU

Top Bookmark and Share
#124721 - 10/21/06 02:10 PM Re: RAM Benchmarks: the effect of 512MB, 1024MB, and 2048MB of memory
Revvin Offline
SimHQ Member

Registered: 04/12/00
Posts: 1612
Loc: United Kingdom
Interesting results though I would back up Gatticus on the difference seen on a friends system load timesfor BF2 though actual FPS increase was hardly noticeable. I've also seen some say WWIIOL runs a lot better once you go beyond 1GB, I'd like to see if thats true.
_________________________
Revvin
The CH Hangar - Profiles, news and help for CH users

Top Bookmark and Share
#124722 - 10/21/06 09:00 PM Re: RAM Benchmarks: the effect of 512MB, 1024MB, and 2048MB of memory
rootango Offline
SimHQ Member

Registered: 12/29/00
Posts: 652
Loc: everywhere
nice article.

i was thinking of increasing my 1gb ram to 2 gb, but might hold off a little longer now.
_________________________
Be advised, we got zips in the wire

Top Bookmark and Share
#124723 - 10/22/06 04:56 AM Re: RAM Benchmarks: the effect of 512MB, 1024MB, and 2048MB of memory
Sepp Offline
SimHQ Member

Registered: 09/09/06
Posts: 494
2Gb of RAM in RO is known as 'Sniper Qualification RAM', because those with it always get into a fresh map first and are able to grab the sniper rifle. \:D

That's the only game I know of where 2 gig of RAM is a known advantage, confirmed by Joe's loading-time tests above. Nice work Joe. \:\)

Top Bookmark and Share
#124724 - 10/22/06 09:23 AM Re: RAM Benchmarks: the effect of 512MB, 1024MB, and 2048MB of memory
Coastie Offline
SimHQ Member

Registered: 04/01/02
Posts: 452
Loc: Va, U.S.
Quote:
Originally posted by WalterNowi:
Joe,
Interesting results. I am wondering what the results would look like for:

2 x 512 MB @1T versus
2 x 1024 MB @1T
I would also like to see this. Especially the 2x1gb test vs 4x512.

Top Bookmark and Share
#124725 - 10/22/06 10:20 PM Re: RAM Benchmarks: the effect of 512MB, 1024MB, and 2048MB of memory
Weasel_Keeper Offline
Asst Forums Manager
SimHQ Senior Member

Registered: 06/28/01
Posts: 7637
Loc: Fort Wayne, Indiana, USA
About the rFactor 1150 load time, I have 2x512 (P4 3.6G) and am seeing anywhere from a minute to 1:20 to load from desktop to main menu. I think it's due more to the sim searching for servers before starting.

Was your test done offline?

_________________________
"Cave Putorium!"
SoWW #2485

Top Bookmark and Share
#124726 - 10/23/06 07:12 AM Re: RAM Benchmarks: the effect of 512MB, 1024MB, and 2048MB of memory
Joe Offline
SimHQ Technology Editor
SimHQ Lifer

Registered: 04/05/02
Posts: 16196
Loc: Somerville, NJ
Quote:
Originally posted by WalterNowi:
I am wondering what the results would look like for:

2 x 512 MB @1T versus
2 x 1024 MB @1T
Sorry, my motherboard (Asrock 939Dual-SATAII) will only run memory at -2T timing.

Quote:
Originally posted by Weasel_Keeper:
About the rFactor 1150 load time, I have 2x512 (P4 3.6G) and am seeing anywhere from a minute to 1:20 to load from desktop to main menu. I think it's due more to the sim searching for servers before starting.

Was your test done offline?

No, my test was done online. AFAIK rFactor doesn't search for servers until you click the "JOIN" button to see the server list. rFactor does check in with the Matchmaker service when it starts, however, so your long load times could be due to this.

Top Bookmark and Share
#124727 - 10/23/06 07:18 AM Re: RAM Benchmarks: the effect of 512MB, 1024MB, and 2048MB of memory
Jedi Master Offline
Entil'zha
SimHQ Big Kahuna

Registered: 02/15/00
Posts: 29341
Loc: Space Coast, USA
I think rFactor's load time, like the SFP series, is also dictated by the number of mods in the game. The more you have, the longer the load time.
Another thing that's not covered here is in some games you use more RAM as time goes on. I'd say a good third of my games will use 1200-1400MB of RAM, like RO. If you have only 1GB, that means swapfile use at some point. Whether reflected in longer loading times once in the game, or pauses/stutters at other points depends on the game situation.

I remember CFS3 is one where each new plane loaded increases the RAM footprint. I have over 100 planes in the game, and if I sit in the main screen, where the pilot lounges by his plane, and just switch from plane to plane to plane, by the time I reach 40 planes or so the game is pushing 2GB RAM used.


The Jedi Master
_________________________
Back off, man. I'm a scientist.

Top Bookmark and Share
#124728 - 10/23/06 12:30 PM Re: RAM Benchmarks: the effect of 512MB, 1024MB, and 2048MB of memory
John Reynolds Offline
Hardware Editor
SimHQ Member

Registered: 01/30/01
Posts: 673
Nice write-up, Joe.
_________________________
Hardware Editor, SimHQ.com

Top Bookmark and Share
#124729 - 10/27/06 09:56 PM Re: RAM Benchmarks: the effect of 512MB, 1024MB, and 2048MB of memory
speedbump Offline
SimHQ Member

Registered: 11/23/05
Posts: 4273
Loc: Corsicana
Interesting. I would think that the load placed on the PSU would be greater for 2G vs 1G even if the 2G was not being fully used? I've read on some forums of some PSUs that were borderline not being able to handle 2G of ram especially if it is 4x512.
_________________________
"Warm winds blowing, heat and blue skies and a road that goes forever"

ASUS P5N-E SLI..E6300 Wolf @ 3.8Ghz..2x1GB DDR2 AxeRam
2x320GB Seagate RAID 0..1TB WD..PNY 8800GT 512..Win7Pro


www.razzledazzleart.com


Top Bookmark and Share
#124730 - 11/24/06 01:51 PM Re: RAM Benchmarks: the effect of 512MB, 1024MB, and 2048MB of memory
Xander Fulton_dup1 Offline
SimHQ Member

Registered: 10/30/06
Posts: 173
Quote:
Originally posted by Joe:

1x512MB (512MB total), 400 MHz, 2.5-3-3-11-2T: 26 seconds
2x512MB (1024MB total), 400 MHz, 2.5-3-3-11-2T: 11 seconds
...
1x512MB (512MB total), 400 MHz, 2.5-3-3-11-2T: 53 seconds
2x512MB (1024MB total), 400 MHz, 2.5-3-3-11-2T: 21 seconds
Errr...most of the tests show a pretty massive jump right here, but...are you SURE it's the 'going from 512mb to 1gb of ram'? I mean, the Opteron you tested with is dual-channel. Going from one stick to two sticks *doubles* the memory bandwidth (IE., 1x512mb PC3200 = 3.2gb/s bandwidth theoretical. 2x512mb PC3200 = 6.4gb/s bandwidth theoretical.)

Surely, that would make an ENORMOUS difference in performance, no? Wouldn't a better test of 512mb have been to do 2x256mb?

Top Bookmark and Share
#124731 - 11/28/06 05:52 PM Re: RAM Benchmarks: the effect of 512MB, 1024MB, and 2048MB of memory
Joe Offline
SimHQ Technology Editor
SimHQ Lifer

Registered: 04/05/02
Posts: 16196
Loc: Somerville, NJ
speedbump,

Compared to a fast CPU and a current- or last-generation video card, my understanding is that memory power consumption isn't too important.


Quote:
Originally posted by Xander Fulton:
Quote:
Originally posted by Joe:
[qb]
1x512MB (512MB total), 400 MHz, 2.5-3-3-11-2T: 26 seconds
2x512MB (1024MB total), 400 MHz, 2.5-3-3-11-2T: 11 seconds
...
1x512MB (512MB total), 400 MHz, 2.5-3-3-11-2T: 53 seconds
2x512MB (1024MB total), 400 MHz, 2.5-3-3-11-2T: 21 seconds
Errr...most of the tests show a pretty massive jump right here, but...are you SURE it's the 'going from 512mb to 1gb of ram'? I mean, the Opteron you tested with is dual-channel. Going from one stick to two sticks *doubles* the memory bandwidth (IE., 1x512mb PC3200 = 3.2gb/s bandwidth theoretical. 2x512mb PC3200 = 6.4gb/s bandwidth theoretical.)
Correct. As stated in the first post, 1x512 is a single channel configuration while 2x and 4x512 are dual channel configurations.

Quote:
Surely, that would make an ENORMOUS difference in performance, no? Wouldn't a better test of 512mb have been to do 2x256mb?
Of course that would have been a better test. There's always a better test. But, I don't have 2x256MB of DDR400 RAM. Come to think of it, I'm not sure I have 2x256MB of any RAM, and if I do it's PC100 or PC133. My point: I conveniently had a decent enough RAM selection matched with a mid-to-high-end machine to run these tests for SimHQ, even if they're not perfect.

Top Bookmark and Share
#124732 - 11/30/06 04:17 PM Re: RAM Benchmarks: the effect of 512MB, 1024MB, and 2048MB of memory
Lamina Offline
SimHQ Member

Registered: 02/12/02
Posts: 265
Loc: Puerto Rico
The improvements from 1GB to 2GB should not be in FPS, but in improved multi tasking (alt-tabbing out of games) reducing stuttering, etc.

Top Bookmark and Share
#124733 - 11/30/06 09:26 PM Re: RAM Benchmarks: the effect of 512MB, 1024MB, and 2048MB of memory
Joe Offline
SimHQ Technology Editor
SimHQ Lifer

Registered: 04/05/02
Posts: 16196
Loc: Somerville, NJ
Multi-tasking I can understand. Stuttering, however is FPS, isn't it? Or am I misunderstanding you?

Top Bookmark and Share
#124734 - 12/02/06 10:31 AM Re: RAM Benchmarks: the effect of 512MB, 1024MB, and 2048MB of memory
senneville Offline
SimHQ Member

Registered: 03/19/06
Posts: 129
Loc: canada
Hi
I have the same board as you the asrock939 and i have 2 512MB OF RAM .I play fsx wich is very much ram hungry,and was wondering if i should get 2 more 512mb or 2 1024 mb and would those last 2 share well with the 512 bars?Dual core is not being used yet i think with games.Does it meen that half the rams are just sitting there waiting for the dual core to be used?I hope you understand my saying im not always clear .Thx for helping .Robert

Top Bookmark and Share
#124735 - 12/02/06 04:49 PM Re: RAM Benchmarks: the effect of 512MB, 1024MB, and 2048MB of memory
Lamina Offline
SimHQ Member

Registered: 02/12/02
Posts: 265
Loc: Puerto Rico
Quote:
Originally posted by Joe:
Multi-tasking I can understand. Stuttering, however is FPS, isn't it? Or am I misunderstanding you?
FPS drops to be exact.

For example, you could be running Battlefield 2 with textures on the highest settings, and have 60fps because you have a nice video card and CPU, but you only 1GB of RAM. You will experience stutters when you move into new areas, the fps will appear to drop to single digits for seconds as the game reads textures from the hard disk because there's not enough RAM available. The game will basically skip because it couldn't access and display the necessary textures in time. It could be so bad that you'd have periodic stutters every few seconds, with excellent framerate in between stutters.

If you are running games like BF2, Red Orchestra, Oblivion, Armed Assault, at resolutions of 1600x1200 with high details, you WILL have stutters unless you move to 1.5GB of RAM minimum. Red Orchestra alone can have a RAM footprint of 1.3GB.

Because of the nature of stutters, they dont show very well on FPS benchmarks. The editor would have to play the game and report on these stutters, the frequency, etc.

Quote:
Originally posted by senneville:
Hi
I have the same board as you the asrock939 and i have 2 512MB OF RAM .I play fsx wich is very much ram hungry,and was wondering if i should get 2 more 512mb or 2 1024 mb and would those last 2 share well with the 512 bars?Dual core is not being used yet i think with games.Does it meen that half the rams are just sitting there waiting for the dual core to be used?I hope you understand my saying im not always clear .Thx for helping .Robert
With FSX I definitely recommend 2GB of RAM, and as much video memory as possible. Preferably 512mb of video ram on your video card. I'm not making this up, one of the FSX devs said it over at the Something Awful forums.

I havent even bothered buying that sim until I receive my Geforce 8800GTS, and that is a lower end card as far as FSX's engine is concerned. That sim wants a video card with 1GB of RAM.

Top Bookmark and Share
#124736 - 12/16/06 01:09 PM Re: RAM Benchmarks: the effect of 512MB, 1024MB, and 2048MB of memory
Cas141 Offline
SimHQ Member

Registered: 12/30/00
Posts: 2745
Loc: Northern hemisphere
This is from Lomac forum-

"Yeah, I know what you mean. I got 4GB of RAM and all my sims/games purr smooth as a kitten. [/QB][/QUOTE]

I've been thinking about more Ram, but I read that 4 gig is OTT, and can be counter-productive?
What did you start with, and did you increase incrementally? If so, was there really a genuine increase in FPS as you did so?
I've got 2 gigs- wondering if i should increase"

It is me asking the last question. What do you guys think?
_________________________
Mankind's problem is not failing to know the difference between right and wrong; - It is failing to know the difference between different and wrong

Top Bookmark and Share
#124737 - 12/17/06 04:06 PM Re: RAM Benchmarks: the effect of 512MB, 1024MB, and 2048MB of memory
Joe Offline
SimHQ Technology Editor
SimHQ Lifer

Registered: 04/05/02
Posts: 16196
Loc: Somerville, NJ
I think the Lock On results answer that question pretty clearly. I saw very little improvement moving from 1GB to 2GB, and I think moving to 4GB is a waste of money.

Top Bookmark and Share
#124738 - 12/27/06 05:31 PM Re: RAM Benchmarks: the effect of 512MB, 1024MB, and 2048MB of memory
Allen Offline
SimHQ Member

Registered: 10/13/99
Posts: 3152
Loc: Ohio USA
Quote:
Originally posted by Joe:
Sorry, my motherboard (Asrock 939Dual-SATAII) will only run memory at -2T timing.

...
I just write FWIW (to anyone interested) to indicate the ASRock motherboard itself does not preclude 1T with 2GB -- mine ran PC3500 at 1T with no tweaking required.

Minimal Details: Corsair DDR TWINX PC3500 2x1GB at 2,3,2,5,1T with a 10 percent overclock of the Opteron CPU and the RAM at 440MHz (~PC3500) -- all other mb settings nominal or slightly overdriven by the 10%, as applicable. Since then, I changed to my current motherboard which runs at the same nominal slightly-overclocked settings at 1T. FWIW.

I assume your mb does not do 1T because, as I remember, your mb is overclocked/tweaked substantially. Also, 1T often only works for two RAM slots and fails for 4 populated RAM slots on most mb -- why I got 2x1GB vs 4x512MB.

And, yes, I got 2GB for Oblivion. It also helps minimize stuttering in Gothic 3. But, most older games were designed around 1GB max I think.
_________________________
ATI HD4870 512MB, ATI650 HD Tuner, Phenom II 940 @3.35GHz, 4x2GB OCZ PC8500, ASUS M3A78-T, WDigital + 4x Seagate 7200.xx = 3TB+, Plextor DVD, PCP&C S75CF 750W, Enermax Full-Twr, 2xSony CRT (24" 16:10 + 21" 4:3), 5.1 Spkr, AT ATH-AD700 Phones, Cougar+FSSB, Saitek X52 Pro, Windows 7 HP 64b, Windows 7 RC 64b, Vista HP 64b

Top Bookmark and Share
#2120210 - 01/22/07 05:53 PM Re: RAM Benchmarks: the effect of 512MB, 1024MB, and 2048MB of memory [Re: Joe]
flanker Offline
SimHQ Member

Registered: 08/07/02
Posts: 253
Thanks, Joe you just saved me $180.00. I was going to add 1024 ram to get to 2048 but not after reading your post.

Top Bookmark and Share
#2204351 - 05/05/07 03:45 PM Re: RAM Benchmarks: the effect of 512MB, 1024MB, and 2048MB of memory [Re: flanker]
norden_x Offline
SimHQ Junior Member

Registered: 05/04/07
Posts: 10
Did anyone mention a that not using as much virtual memory is one reason loading may be faster with 1024 over 512? (not just the doubling in bandwidth with dual channel)

Top Bookmark and Share
#2206402 - 05/08/07 09:43 AM Re: RAM Benchmarks: the effect of 512MB, 1024MB, and 2048MB of memory [Re: norden_x]
Joe Offline
SimHQ Technology Editor
SimHQ Lifer

Registered: 04/05/02
Posts: 16196
Loc: Somerville, NJ
Norden_x,

I think you mean the page file, as virtual memory is defined as all physical memory (RAM + page file) available to the OS.

I think that not using the page file as much with 1024MB of memory has a much larger effect than the jump from single- to dual-channel does. There is clearly less hard drive activity with 1024MB.

Top Bookmark and Share
#2215663 - 05/20/07 01:31 PM Re: RAM Benchmarks: the effect of 512MB, 1024MB, and 2048MB of memory [Re: Joe]
NH2112 Online   content
SimHQ Member

Registered: 06/06/06
Posts: 3840
Loc: Windham ME
My current system is a 3.0GHz P4 (OCd to 3.3GHz for now) with 2x256MB DDR400, CAS latency 2.5. I want to upgrade but don't really know how much I'd gain from going to 1.5GB or 2GB as opposed to just adding 2x256MB. Crucial shows these upgrades:

http://www.crucial.com/systemscanner/viewscanbyid.aspx?id=81D92C4F31D926C3&tabid=&pagename=


I'm planning on flying some sims on this PC, most likely Falcon 4.0 and Jane's F15. I know my current configuration is fine for F15, but would 4x256MB and a 128MB GeForce 5200FX be enough for F4?
_________________________
Phil

Hey politician, can't believe a word you say
Almighty media, whose truth d'you sell today?
Watchdog of justice, who keeps their eye on you?
Con man, song in hand, who you singin' to?
The more I get to see, the less I understand
I'm just another ordinary man.

Top Bookmark and Share
#2215686 - 05/20/07 02:11 PM Re: RAM Benchmarks: the effect of 512MB, 1024MB, and 2048MB of memory [Re: NH2112]
ted3 Offline
SimHQ Member

Registered: 05/11/07
Posts: 165
Memory is cheap now, just as well get a 2x512MB kit.

F4 isnt very demanding on hardware so 1GB total should be enough. In the webshops i buy hardware from 2x256 is higher price per MB.

Top Bookmark and Share
#2219072 - 05/24/07 11:03 AM Re: RAM Benchmarks: the effect of 512MB, 1024MB, and 2048MB of memory [Re: ted3]
Joe Offline
SimHQ Technology Editor
SimHQ Lifer

Registered: 04/05/02
Posts: 16196
Loc: Somerville, NJ
Your GPU is more more of a limiting factor than your memory if you have an FX5200.

Top Bookmark and Share
#2219597 - 05/24/07 10:09 PM Re: RAM Benchmarks: the effect of 512MB, 1024MB, and 2048MB of memory [Re: Joe]
NH2112 Online   content
SimHQ Member

Registered: 06/06/06
Posts: 3840
Loc: Windham ME
Yeah, I'm thinking of hitting eBay for an AGP X850XT or something similar. I don't want to dump much more than $100 or so into it because I'm planning on starting to gather parts for a new build soon.
_________________________
Phil

Hey politician, can't believe a word you say
Almighty media, whose truth d'you sell today?
Watchdog of justice, who keeps their eye on you?
Con man, song in hand, who you singin' to?
The more I get to see, the less I understand
I'm just another ordinary man.

Top Bookmark and Share
#2226915 - 06/03/07 03:33 PM Re: RAM Benchmarks: the effect of 512MB, 1024MB, and 2048MB of memory [Re: NH2112]
Beltfed Offline
SimHQ Member

Registered: 02/29/00
Posts: 1106
Loc: NYC, NY, USA
SO here's an interesting article about RAM and the 4gb barrier. Pretty cool.


http://www.dansdata.com/askdan00015.htm
_________________________
Tim "Beltfed" Harrison
NCOIC,
718th TFW,
SkyScrapers
http://www.718tfw.com

Top Bookmark and Share
#2426907 - 01/19/08 06:28 PM Re: RAM Benchmarks: the effect of 512MB, 1024MB, and 2048MB of memory [Re: Beltfed]
The Nephilim Offline
TAW 2.0 Beta Tester
SimHQ Member

Registered: 01/26/02
Posts: 2370
Loc: 3rd Stone from the Sun !!
I am currently Building a new computer with slightly older components like the PIV 570j CPU and a Geforce 7950GT PCIe. The reason, I use the 3D glasses and the newer cpus and vid cards are Not supported with the new drivers and the old ones dont work. Bummer but nVidia is to be working on compatible Drivers. But luckily the board I am getting is compatible with the newer CPU's up to 1333FSB well when they come out anyhow ;\)

I have bought 2 gb memory as I have dual Channel capable Mobo, I mostly Fly the Older Sims like F4:AF-OF JF/A-18 LOMAC 1.12 TAW. Would getting 1 more gig do anything. I also have Doom3 and Quake 4 that I will Probally be replaying but as for the newer FPS's I have No interest. I dont see any newer Demanding Flight sims coming anytime soon except maybe BlackShark or what ever they are calling it these days!!


Edited by The Nephilim (01/19/08 06:29 PM)
_________________________
Intel PIV 3.8Ghz HT / Asus Striker II Formula / OCZ 4GB Memory / 2 BFG GeForce 7950GT's OC SLI / Audigy 2 ZS / Sharp DLP XR-10XL / 45" Screen / X52 HOTAS /

Top Bookmark and Share
#2436830 - 01/31/08 08:38 AM Re: RAM Benchmarks: the effect of 512MB, 1024MB, and 2048MB of memory [Re: The Nephilim]
Joe Offline
SimHQ Technology Editor
SimHQ Lifer

Registered: 04/05/02
Posts: 16196
Loc: Somerville, NJ
2GB will be fine.

Top Bookmark and Share
#2538210 - 06/23/08 05:47 PM Re: RAM Benchmarks: the effect of 512MB, 1024MB, and 2048MB of memory [Re: Joe]
The Nephilim Offline
TAW 2.0 Beta Tester
SimHQ Member

Registered: 01/26/02
Posts: 2370
Loc: 3rd Stone from the Sun !!
Well the 2GB Worked rather Well. I ended up buying Oblivion and GTR2 they seem to use alot of RAM and It works rather well even on an older parts computer. well the really old part is the Geforce 7950GT and Audigy 2 zs but they are holding up rather well. Even using the Older Drives.

I do Alot of 3D gaming with EDimensionals 3D Glasses and I know doing 3D Needs Alot of memory as it does a double image and the textures are twice that than a single image. I wonder IF I would Benefit from 4GB Memory as I do 3D Gaming would the added memory be more effective for 3D Stereo??

Note the Note:













I also was reading about Virtual memory at Microsoft here is the Link. I did setup the Virtual memory as they described but How would I tell if it actually is working or effective??



http://support.microsoft.com/kb/314482/EN-US/


Edited by The Nephilim (06/23/08 06:08 PM)
_________________________
Intel PIV 3.8Ghz HT / Asus Striker II Formula / OCZ 4GB Memory / 2 BFG GeForce 7950GT's OC SLI / Audigy 2 ZS / Sharp DLP XR-10XL / 45" Screen / X52 HOTAS /

Top Bookmark and Share
#2540340 - 06/27/08 04:26 AM Re: RAM Benchmarks: the effect of 512MB, 1024MB, and 2048MB of memory [Re: The Nephilim]
Phoenix Offline
SimHQ Member

Registered: 05/08/05
Posts: 554
It's not that you really need four gigs for what you stated, but if you can invest in it cheaply then I would go with it. It would benefit you if you plan to run newer 3d software/games and if you have an x64 OS or if you have Vista. 4GB would help in the following was:

1) In the newest software (like Crysis), it would *significantly* cut down on load time.
2) Changes in average FPS would be little to nil, but you would also see a significant rise in minimum FPS.
3) As a general rule, you want to use the same kind of memory modules. If you upgrade now you can get the same ones. If you wait until later, your current ones might be out of stock and you may have to mix and match -- which from my experience, can cause problems.

Ways that 4GB might not help

1) You may not be able to overclock your memory as much
2) If you use 4 sticks, your memory has less breathing space and can get very hot.
3) Unless you plan to run the most cutting edge software, 4GB will not help you all that much

Top Bookmark and Share
#2541429 - 06/28/08 04:03 PM Re: RAM Benchmarks: the effect of 512MB, 1024MB, and 2048MB of memory [Re: Phoenix]
The Nephilim Offline
TAW 2.0 Beta Tester
SimHQ Member

Registered: 01/26/02
Posts: 2370
Loc: 3rd Stone from the Sun !!
I kinda Figured 4 Gigs would be kinda overkill as I Mostly Play the Older Sims F4:AF/OF, JF/A-18, LOMAC:FC, EECH, EE2. I also Bought Oblivion and GTR2 which works rather well on the older system, I was surprised I could run Oblivion an High settings at 1024x768 ;\)

Well what about the Virtual Memory as Described by Microsoft in my last part of the Thread?? How can I tell if that is working and is it effective??

I have a SATA 500gb drive and I put in a 2nd drive that is IDE 200gb and use that for the Virtual Memory. I have NOT Partitioned it I will do that ltr. but for now I just want to know if using a IDE drive is effective for Vitrual memory as described in the bleow Link. I adjusted the Virtual memory as described in the link 1.5 to 3 times my RAM.


http://support.microsoft.com/kb/314482/EN-US/
_________________________
Intel PIV 3.8Ghz HT / Asus Striker II Formula / OCZ 4GB Memory / 2 BFG GeForce 7950GT's OC SLI / Audigy 2 ZS / Sharp DLP XR-10XL / 45" Screen / X52 HOTAS /

Top Bookmark and Share
#2544024 - 07/02/08 06:53 PM Re: RAM Benchmarks: the effect of 512MB, 1024MB, and 2048MB of memory [Re: The Nephilim]
Phoenix Offline
SimHQ Member

Registered: 05/08/05
Posts: 554
But is it better to put your pagefile on your fast master SATA, or on your significantly slower (but less used) IDE?

Top Bookmark and Share
#2544066 - 07/02/08 08:15 PM Re: RAM Benchmarks: the effect of 512MB, 1024MB, and 2048MB of memory [Re: Phoenix]
The Nephilim Offline
TAW 2.0 Beta Tester
SimHQ Member

Registered: 01/26/02
Posts: 2370
Loc: 3rd Stone from the Sun !!
that is the question ;\)

In the Article it did say to put tut the pagefile on a HD other than the one you are using for the Programs so I figured use the IDE??
_________________________
Intel PIV 3.8Ghz HT / Asus Striker II Formula / OCZ 4GB Memory / 2 BFG GeForce 7950GT's OC SLI / Audigy 2 ZS / Sharp DLP XR-10XL / 45" Screen / X52 HOTAS /

Top Bookmark and Share
#2544169 - 07/02/08 11:57 PM Re: RAM Benchmarks: the effect of 512MB, 1024MB, and 2048MB of memory [Re: The Nephilim]
Rilex Offline
SimHQ Member

Registered: 02/20/02
Posts: 2918
Loc: Washington
 Originally Posted By: The Nephilim

Well what about the Virtual Memory as Described by Microsoft in my last part of the Thread?? How can I tell if that is working and is it effective??


Don't worry about page file placement. If you were significantly paging in to or out of the page file while playing games, it'd be a slide show. This setting is not meant for gamers. It is meant for people running SQL Server, Virtual Server, and so forth.

The answer to hard paging (paging in and out of backing store, be it the page file or application binaries) is one thing and one thing only: add more RAM.

Top Bookmark and Share
#2545215 - 07/04/08 05:22 PM Re: RAM Benchmarks: the effect of 512MB, 1024MB, and 2048MB of memory [Re: Rilex]
The Nephilim Offline
TAW 2.0 Beta Tester
SimHQ Member

Registered: 01/26/02
Posts: 2370
Loc: 3rd Stone from the Sun !!
Hey that is Great Info THNX Rilex. So can I just let windows manage my Virtual memory??

Will XP Pro 32bit handle 4 gigs??
_________________________
Intel PIV 3.8Ghz HT / Asus Striker II Formula / OCZ 4GB Memory / 2 BFG GeForce 7950GT's OC SLI / Audigy 2 ZS / Sharp DLP XR-10XL / 45" Screen / X52 HOTAS /

Top Bookmark and Share
#2545241 - 07/04/08 06:47 PM Re: RAM Benchmarks: the effect of 512MB, 1024MB, and 2048MB of memory [Re: The Nephilim]
Rilex Offline
SimHQ Member

Registered: 02/20/02
Posts: 2918
Loc: Washington
1) Yes
2) Kinda. The problem is that the with x86 and boards which do not have the ability to map addresses >4G, they will map their hardware addresses starting at 4G on down until it is satisified (so this means addresses for the BIOS, any video card memory, etc.). Windows XP/Vista, as of XP SP2 x86, is then further artifically limited to addressing 3.5GB of physical memory. Of course, unless you have very little hardware, you may not get that much due to the first issue with the BIOS mapping addresses in the 4G range. I was able to get 3.5GB by setting my AGP aperature to 64MB (I had a GeForce 6800GT 256MB, IIRC).

I'm running Vista x64 w/ 8GB RAM now, though ;\)

Top Bookmark and Share
#2545728 - 07/06/08 12:39 AM Re: RAM Benchmarks: the effect of 512MB, 1024MB, and 2048MB of memory [Re: Rilex]
Phoenix Offline
SimHQ Member

Registered: 05/08/05
Posts: 554
Good grief...8GB? Are you running a server?

Top Bookmark and Share
#2545991 - 07/06/08 01:40 PM Re: RAM Benchmarks: the effect of 512MB, 1024MB, and 2048MB of memory [Re: Phoenix]
Rilex Offline
SimHQ Member

Registered: 02/20/02
Posts: 2918
Loc: Washington
Usually 3 virtual machines at all times.

Top Bookmark and Share
#2570184 - 08/17/08 10:37 PM Re: RAM Benchmarks: the effect of 512MB, 1024MB, and 2048MB of memory [Re: Rilex]
Rilex Offline
SimHQ Member

Registered: 02/20/02
Posts: 2918
Loc: Washington
Joe, you should update this with:

4x1
2x2
4x2
2x4 (if you can afford it)

Top Bookmark and Share
Page 1 of 5 1 2 3 4 5 >



Forum Use Agreement | Privacy Statement | SimHQ Staff
Copyright 2009, SimHQ Inc. All Rights Reserved.