Sign in with
Sign up | Sign in
Your question
Closed

Exploring SSD Performance In Battlefield 3, F1 2011, And Rift

Last response: in Reviews comments
Share
November 25, 2011 3:14:11 AM

Interesting to see it won't imrpove all gameplay
Score
8
November 25, 2011 3:44:36 AM

Makes sense. I'll install sw:tor to the ssd after i figure out how to make its 20GB fit on it lol... the rest go to the HDD...

Good read. Thanks for being so thorough.
Score
3
Anonymous
November 25, 2011 3:46:02 AM

IMO, a WD Caviar Black or a Samsung Spinpoint F3 would have been a better test drive than the Green one.
Score
26
November 25, 2011 3:50:57 AM

Thanks to all the reviews you made here guys. A couple of months ago I started planning my ssd purchase, and i decided to buy a Kingston HyperX 240gb bundle kit. I cant wait till it arrives. I have advanced computer skills, so many things I thought from personal experiences are published here, like I knew some games are more write dependant or read dependant from and HDD, and about the apps on the background on WIN7, and many many more. I already ordered my SSD, but its nice to have a serious review about ssds on gaming performance. Its all about the speed that the apps can be done by write/read on a ssd on the background that really matters, from this simple thing is that anyone willing to upgrade to an ssd can benefit with a smoother playable experience.
Score
0
November 25, 2011 4:07:59 AM

Storage can actually make a difference in FPS in certain situations. Only put in 2gb or so of RAM in the test system and rerun Battlefield 3, and the difference in frame rates will be different between HDD and SSD as the game switches to secondary storage once RAM is exhausted.
Score
-7
November 25, 2011 4:20:47 AM

True, but it is much more cost effective to add more RAM to your system in that circumstance than to get an SSD large enough for your OS and a couple of games. With RAM so inexpensive these days there is absolutely no reason to be running low on it under any gaming scenario.
Score
19
November 25, 2011 4:22:10 AM

Agreed with xyzqwerty, please do another test with faster drives such as wd black or the f3. Comparing the ssd with green version is like racing your turtle with the neighbor's hare.
Score
4
November 25, 2011 5:09:03 AM

cumi2k4Agreed with xyzqwerty, please do another test with faster drives such as wd black or the f3. Comparing the ssd with green version is like racing your turtle with the neighbor's hare.


I'm a little confused why you would want to see that comparison. We established there was no diff between a slow HDD and a fast SSD. And you expect a difference between a fast HDD and fast SSD?

Cheers,
Andrew Ku
TomsHardware.com
Score
9
November 25, 2011 5:36:56 AM

asnorton44Interesting to see it won't imrpove all gameplay

I see what you did there. :p 
Score
2
November 25, 2011 6:09:56 AM

acku said:
I'm a little confused why you would want to see that comparison. We established there was no diff between a slow HDD and a fast SSD. And you expect a difference between a fast HDD and fast SSD?

Cheers,
Andrew Ku
TomsHardware.com



He probably want to see that comparison on the "Hard Drive Performance Comparison" page, where is a big difference between SSD and HDD performance (the HDD up to 477% slower)
Score
4
November 25, 2011 6:48:46 AM

Quote:
IMO, a WD Caviar Black or a Samsung Spinpoint F3 would have been a better test drive than the Green one.

The point of using a slower drive than those is to emphasize the effects of using SSD vs HDD when it comes to gameplay/FPS. Of course, if the slower drive is quite on par with SSD on FPS ratings then the faster HDDs will do well too.
Score
-2
November 25, 2011 6:53:30 AM

gmcizzle said:
Storage can actually make a difference in FPS in certain situations. Only put in 2gb or so of RAM in the test system and rerun Battlefield 3, and the difference in frame rates will be different between HDD and SSD as the game switches to secondary storage once RAM is exhausted.


And who the hell would buy an expensive SSD but cannot even put a decent amount of RAM on his system? :pfff: 
Score
14
November 25, 2011 7:09:13 AM

Quite surprised to see Skyrim and WoW left out of the pool. In my opinion games like these give a lot of improvement in SSD. Just think how many times you load in BF3 and compare that with skyrim. Pretty insane :D 

Notably, my laptop with a Intel 510 250GB, my in-game load times are about 2 seconds in skyrim compared to the desktop running 1TB spinpoint at raid takes 4-5 seconds. The same desktop with a crucial M4 takes around 2 seconds. Moreover I never had any spikes/lags with SSDs while gaming.



Score
6
November 25, 2011 7:37:23 AM

masterjawAnd who the hell would buy an expensive SSD but cannot even put a decent amount of RAM on his system?

Nobody would of course, but the point is it can make a difference.
Score
-3
November 25, 2011 8:01:04 AM

ackuI'm a little confused why you would want to see that comparison. We established there was no diff between a slow HDD and a fast SSD. And you expect a difference between a fast HDD and fast SSD?Cheers,Andrew KuTomsHardware.com


If you are just looking at frame rate performance, you are of course correct Andrew; what's the point?. However, what about the launching of the games and level loading? In some games there is a marked difference.

The comparisons I'd like to see would be the load times for game launching and levels with faster hard drive solutions. As an MMORPG player, the zoning times would be of interest - how long does it take to zone into another instance, and does that significantly improve with an SSD?

I'd like to see 2 hard drives in RAID 0 compared. Other single hard drives would be interesting (being a great fan of the Velociraptor series, I'm always interested in those, even if they are kind of an odd duck at their price in today's world) but to a lesser degree for me.

The other test comparisons I'd like to see would be SSDs used in caching systems, such as what the Z68 chipset allows, (a whole article could be done on the effects of size of SSD on caching, I think) and solutions like the OCZ Synapse, and RAM drives. (RAM is so cheap now popping an extra 12GB into a 24GB capable mainboard and setting up a 16GB RAM drive looks like a very attractive way for me to boost performance in my aging X58 system. :)  )

Please don't take any of my suggestions as complaints or that I'm saying your tests were incomplete; I'm just throwing some ideas out. I'm well aware of how much effort goes into producing these articles, and am grateful for all that you do.

;) 


Score
10
November 25, 2011 8:05:38 AM

What happens if you limit the SSD to a SATA 3Gb/s port?
Score
-2
November 25, 2011 8:08:25 AM

I want to see how RAID of HDD compares to single SDD at comparative cost, and not only for games.
Score
3
November 25, 2011 8:28:02 AM

There a lot of things to consider, however smoothness in gaming is still an expensive feature. In the future we´ll be looking at SSDs prices like we do today with RAM.
Score
2
November 25, 2011 8:32:48 AM

perfectblueQuite surprised to see Skyrim and WoW left out of the pool. In my opinion games like these give a lot of improvement in SSD. Just think how many times you load in BF3 and compare that with skyrim. Pretty insane Notably, my laptop with a Intel 510 250GB, my in-game load times are about 2 seconds in skyrim compared to the desktop running 1TB spinpoint at raid takes 4-5 seconds. The same desktop with a crucial M4 takes around 2 seconds. Moreover I never had any spikes/lags with SSDs while gaming.



He did WoW in a previous article.

I would like to have seen Skyrim too, but really Andrew's test suite should be about displaying different facets of storage solution performance before it's about pleasing the readers with games they want to see - and the games he chose are extremely popular. In my opinion, Rift isn't "the latest craze" in MMOGs, it's kind of old news, but it is still a good test to run at least once in that we see how it runs and how it is different than, say, WoW, or the other non-MMOG games.

One thing is very clear, games aren't all the same; not even games of the same type. One of the reasons I kept asking for WoW to be used as a bench in the past was because it was so CPU intensive, and most game weren't. In Cataclysm, that changed, and it became a game that could stress both CPU AND GPU, so I think it is a very good game to use for benches. However, MMOGs are harder to bench with than other games.

You can only use so many games for benchmarks, and some people are going to wish other games were used so an article relates to them more directly. I think it's good to make suggestions, but have trouble understanding why people get surprised when a game isn't used, talk in terms of something being "left out", or seem to take things personally.

;) 
Score
0
November 25, 2011 8:38:38 AM

Arma 2 is one of those games that really leans on hard drive performance. Because it uses a very large continuous world the game struggles to load the textures when travelling any faster than walking speed. There is never any texture/world popping with a Vertex 2 however, even an Intel X25-M is fast enough to solve the problem.

It matters a lot when using a helicopter or plane as trees are notoriously appearing at very short range all of a sudden. While it doesn't affect the frame rate the game works a whole lot better with enough IO performance.
Score
3
November 25, 2011 8:46:43 AM

Wow benefits from an SSD in a few circumstances. With a normal HDD in Dalaran players will pop in very close to the player. An SSD on the other hand ensures the players are loaded much further away and the odd popping never happens. The SSD also seems to load faster between areas, twice as fast or so. However I have never seen Wow suffer frame rate issues related to its continuous use of the drive, and the other benefits are of marginal utility.
Score
2
November 25, 2011 9:55:30 AM

Something is wrong, battlefield loadtime level 1:50 with ssd ?
Score
-2
November 25, 2011 10:05:22 AM

I noticed a column called "disk queue" in the Disk section of perfmon, and i noticed origin and ie launched with a disk queue of 0.xx, while bf3 launched with with a queue of 1.xx. I'm using a 7200 RPM drive.

Question is, how can the disk queue be fractional?
Score
1
November 25, 2011 10:33:24 AM

Is the BF3 launch chart correct? you've listed:

Quote:
I/O Trends:

93% of all operations occur at a queue depth of one
53% of all operations are random
42% of all operations are 4 KB in transfer size
16% of all operations are 8 KB in transfer size
11% of all operations are 16 KB in transfer size
15% of all operations are 32 KB in transfer size
11% of all operations are 128 KB in transfer size


But the chart shows 40% for 8KB.

Also, does
Quote:
average transfer size * IOPS = MB/s

(from Understanding Storage Performance)

still hold in this case?
Score
-2
November 25, 2011 10:52:59 AM

excellent! how easy solution for games
Score
-1
November 25, 2011 12:08:06 PM

am i mistaken or the benchmark for Battlefield 3 only was done on missions ? not multiplayer matches...
I wonder how much different the data would be on the second case.
from a personal experience, my spec is : Corsair Obsidian 650D case|Corsair Hydro H80 Cooler|Asus P8Z68-V Intel Z68|Intel Core i7-2600K @3.40GHz|Kingston HyperX Genesis Grey 8GB CAS 9 1600MHz DualChann|OCZ Vertex 3 120GB|Western Digital Caviar Blue 500GB|Corsair Pro Series Gold 850W modular and i can tell you im always the 1st among my group of friends to be already in the map while all the others using magnetic disk are still loading level... i suppose if you can afford it it just saves you to have to wait ages to load levels. Also in Rift, incredibly fast, didnt even have time to finish the loading screen, i was in after 10%. Definitely recomended to upgrade to SSD for gamers. Keep in mind though that it doesnt affect all the gameplay, like Fps etc.
Score
1
November 25, 2011 1:05:07 PM

uruquioraam i mistaken or the benchmark for Battlefield 3 only was done on missions ? not multiplayer matches...I wonder how much different the data would be on the second case.

Would be similar, it still takes over a minute for me to load a multi level on a HDD, so an SSD should make a difference...Though while actually playing a multiplayer game, there should not be much difference between a HDD's and SSD's performance, assuming most of the read/writes are server side.

However, i also I know that there's a period of heightened HDD activity after i exit the game and return to battlelog.
Score
-1
November 25, 2011 1:58:48 PM

The only issue I have with SSD's is the size, but I have found a partial workaround. I use Steam, which likes to install all games in the same place. For games that don't require the same level of performance with level loads and start-up I use junctions to move them to a conventional drive. Junction Link Magic has done well for me so far, but there are several options. Steam doesn't know the difference and I get to keep Steam and my "high-priority" games on a dedicated SSD with the OS on another.
Score
1
November 25, 2011 2:13:17 PM

Gamer-girl said:
What happens if you limit the SSD to a SATA 3Gb/s port?


Largely you get the same benefit. If OCZ is to be believed, their Vertex 3 speeds go from 525mb/s to around 260mb/s, but IOPs shouldn't really change.

Also today is a good day to buy an SSD. I'm looking at upgrading to a 120gb since there's some over $100 off with the MIR.

Just gotta decide... Force Series 3, Force Series GT, or Vertex 3...
Score
1
November 25, 2011 3:15:20 PM

srgessSomething is wrong, battlefield loadtime level 1:50 with ssd ?


Read page 12.
Score
1
November 25, 2011 3:18:43 PM

ojasIs the BF3 launch chart correct? you've listed:But the chart shows 40% for 8KB. Also, does (from Understanding Storage Performance)still hold in this case?


http://www.tomshardware.com/reviews/ssd-gaming-performa...
512 bytes per sector, yielding 4 KB
Yeah so its the same as the previous article.
Score
1
November 25, 2011 4:27:32 PM

This is the wrong GPU analysis. The right one is here:

http://techreport.com/articles.x/22048/4

Analysis frame by frame.

I don't care if the card does 60 fps when it means half a second at 120 fps, and half a second freezed by disk access.

Far cry 2 was unbearable with HD, because it was loading textures all the time. With SD is another history, you can download large files from internet, run a virus scan, and no problem.
Score
1
November 25, 2011 4:35:19 PM


Question for Andrew Ku.

Great article, but how do you measure the I/O created by the games you profiled in this article ?

Is it using Sysinternal Procmon, hiOmon, Xperf, Perfmon disk counters ? So or something else. I'm curious because I'd like to profile a few applications on my on PC.

(I've looked for the details in this article and august one but have not found the details, other than the reference to Intel IPEAK which only works in XP x86.)

Any clues you can provide would be much appreciated
Score
1
November 25, 2011 5:17:01 PM

marraco said:
This is the wrong GPU analysis. The right one is here:

http://techreport.com/articles.x/22048/4

Analysis frame by frame.

I don't care if the card does 60 fps when it means half a second at 120 fps, and half a second freezed by disk access.

Far cry 2 was unbearable with HD, because it was loading textures all the time. With SD is another history, you can download large files from internet, run a virus scan, and no problem.


Errr.... I don't think we ever claimed this to be a GPU analysis. Second, Don has a more through writeup on Skyrim comparing graphic cards. http://www.tomshardware.com/reviews/skyrim-performance-...

Score
-1
November 25, 2011 5:18:12 PM

watts_354 said:
Question for Andrew Ku.

Great article, but how do you measure the I/O created by the games you profiled in this article ?

Is it using Sysinternal Procmon, hiOmon, Xperf, Perfmon disk counters ? So or something else. I'm curious because I'd like to profile a few applications on my on PC.

(I've looked for the details in this article and august one but have not found the details, other than the reference to Intel IPEAK which only works in XP x86.)

Any clues you can provide would be much appreciated


http://www.tomshardware.com/reviews/ssd-gaming-performa... That explains it all. And the new IPEAK works in x64.
Score
1
November 25, 2011 5:20:15 PM

ojas said:
Would be similar, it still takes over a minute for me to load a multi level on a HDD, so an SSD should make a difference...Though while actually playing a multiplayer game, there should not be much difference between a HDD's and SSD's performance, assuming most of the read/writes are server side.

However, i also I know that there's a period of heightened HDD activity after i exit the game and return to battlelog.


Ojas already covered it. But yes, our trace was based on a campaign mission. However, I did spend time tracing multiplayer games. Overall, I'd say its the same when it comes to storage I/O and memory management.
Score
1
November 25, 2011 6:10:58 PM

I am surprised that there was no measurement and comparison of "Minimum" frame rate between HDD and SSD. I was of the understanding that the "dips" one sees in play as a result of the slower response time in magnetic storage, is quantifiable as what the minimum framerate is (as opposed to using a youtube vid as evidence)
What I primary based my SSD purchase on, was reducing the amount of time that the game spends under 30fps!
Score
0
November 25, 2011 6:51:25 PM

ackuhttp://www.tomshardware.com/review [...] ,2991.html That explains it all. And the new IPEAK works in x64.


Hi Acku, I read that article a few time, and even posted the same question last august, What did Andrew Ku use, the last "official" version of the Intel IPEAK was 3.0 - 8/31/1999. The old Intel website is down, http://developer.intel.com/design/ipeak.
So "Ipeak version "v5.2" does not exist. If it does it would be nice to have the reference to it.

Did a private party acquire the source code and continue it's development as a new product ?
Score
1
November 25, 2011 7:07:40 PM

watts_354Hi Acku, I read that article a few time, and even posted the same question last august, What did Andrew Ku use, the last "official" version of the Intel IPEAK was 3.0 - 8/31/1999. The old Intel website is down, http://developer.intel.com/design/ipeak. So "Ipeak version "v5.2" does not exist. If it does it would be nice to have the reference to it.Did a private party acquire the source code and continue it's development as a new product ?

It is only available directly through Intel. We do not have permission to distribute it. Sorry.

Other companies have licensed the code base, as I understand it. PCMark 7 and Vantage both use IPEAK as the basis of their HDD test suite.

Cheers,
Andrew Ku
TomsHardware.com
Score
1
November 25, 2011 8:40:06 PM

how many times does it have to be said.
resolution @ 1920 x 1080 should be the standard, nothing less...
Score
0
November 25, 2011 8:43:43 PM

verbalizer said:
how many times does it have to be said.
resolution @ 1920 x 1080 should be the standard, nothing less...


Please read page 12 and 13. We use 1920 x 1080 in our ssd vs hdd comparison.
Score
1
November 25, 2011 8:47:18 PM

acku said:
Please read page 12 and 13. We use 1920 x 1080 in our ssd vs hdd comparison.

much appreciated.
it should always be on every testing set-up as well.
Score
-1
November 25, 2011 8:53:00 PM

BrightCandleWow benefits from an SSD in a few circumstances. With a normal HDD in Dalaran players will pop in very close to the player. An SSD on the other hand ensures the players are loaded much further away and the odd popping never happens. The SSD also seems to load faster between areas, twice as fast or so. However I have never seen Wow suffer frame rate issues related to its continuous use of the drive, and the other benefits are of marginal utility.


With WoW a couple of Vertex 2s in RAID-0 loads instances very very fast. However I still see plenty of "popping" of monsters when flying in quickly to a heavily populated zone. Distance at which monsters appear is server controlled and thus "popping" can be the server's fault. At least with an SSD though, you know that its not your machine's fault for being slow to load textures.
Score
1
November 25, 2011 10:17:01 PM

verbalizer said:
much appreciated.
it should always be on every testing set-up as well.


Well, it doesn't change the fundamentals of our trace. It's only going to scale the numbers ever so slightly higher. The ratio and percentages will remain the same.

I used 1680x1050 because it allows me to monitor IO in real-time as I'm playing the game, as well as keep an eye on CPU and RAM utilization. The last is important because it helps explain if a game is loading into RAM or caching onto the disk.
Score
2
November 25, 2011 10:24:04 PM

I for one am so glad that I found this site in the last year. I went from a laptop which i had been using to picking components from scratch to build up a desktop. I found this site in the process of relearning about computer parts and getting up to speed on the new tech. I mainly play games on my pc but occasionally do other stuff and reviews such as this one are welcome in my book. I read most every review that Tom's does, and reviews such as this are a bit hard to come by. I would love to see more of this, keep it up.
Score
3
November 26, 2011 3:23:40 AM

ackuhttp://www.tomshardware.com/review [...] 991-4.html512 bytes per sector, yielding 4 KBYeah so its the same as the previous article.

Ah, silly me, i didn't see that the chart said sectors!! Thanks for pointing that out...

So yeah if a vertex 3 does about 60k IOPS ideally, lets assume for a moment that it's not ideal and does 50k IOPS.
SO now, do the 128KB transfers get transferred at 128*50,000=7,680,000 KB/s = 7500 MB/s???
That's hardly possible, but 4KB transfers at 50k IOPS at 195 MB/s is a fair calculation.

Or are IOPS strictly for an average? It's a bit confusing how to use them.
Score
1
November 26, 2011 8:36:24 AM

Id have to say that the biggest benefit I see when using my ssd is that when the game locks up or is loading it loads a few seconds faster. Ive seen youtube videos of guys loading games on their xbox or pc with normal hard drives in them and Im stunned at how long the loading screens are compared to mine. Like loading a game of mass effect on xbox... when the little movie for the load screen comes up I expect it to be gone in about 5 seconds... on xbox it takes fully 10-15 sec. I actually exclaimed "Come on already while Im young" at the video... While playing BF3 Im almost always the first guy to load the level and be ready... doesnt help though cause you have to wait 30 seconds for everyone else to choose their stuff and get ready too. It doesnt help at all when it comes to frame rates but it does help when the games crash :D  I can get back in game ten times quicker than with a normal hard drive... great use of 250 bucks huh?
Score
1
November 26, 2011 8:53:51 AM

ojasAh, silly me, i didn't see that the chart said sectors!! Thanks for pointing that out...So yeah if a vertex 3 does about 60k IOPS ideally, lets assume for a moment that it's not ideal and does 50k IOPS.SO now, do the 128KB transfers get transferred at 128*50,000=7,680,000 KB/s = 7500 MB/s???That's hardly possible, but 4KB transfers at 50k IOPS at 195 MB/s is a fair calculation.Or are IOPS strictly for an average? It's a bit confusing how to use them.


Ah. I see the confusion. (Pulls out glasses and queue school house rock music....) :p 

IOPS * TransferSizeInBytes = BytesPerSec

Remember your units. Second to get IOPS, you'd have to do the weighted average of the transfer sizes. In the scenario of 93% being 4 KB, 2 % being 128 KB, etc.. etc...
Score
2
November 26, 2011 8:57:38 AM

adjmanI for one am so glad that I found this site in the last year. I went from a laptop which i had been using to picking components from scratch to build up a desktop. I found this site in the process of relearning about computer parts and getting up to speed on the new tech. I mainly play games on my pc but occasionally do other stuff and reviews such as this one are welcome in my book. I read most every review that Tom's does, and reviews such as this are a bit hard to come by. I would love to see more of this, keep it up.


I'm glad you liked it. Honestly, sometimes these traces have me scratching my head a bit. For example, I expected more games to be similar than different. That's actually turning out not to be the case, and it should have dawned on me earlier since I'm a programmer.

What's interesting is that intensive graphics =! a lot of benefit with an SSD. There are games like WoW and Rift that are clearly random and generic looking to the more impressive Battlefield 3, which is basically all sequential reads.

Cheers,
Andrew Ku
TomsHardware.com
Score
1
!