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

Page 3 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Toggle sidebar Toggle sidebar
Status
Not open for further replies.
[citation][nom]gmcizzle[/nom]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.[/citation]

Not really true. Some games only cache to RAM. i.e. BF3, if space is unavailable the game usually dumps clears some data to make new space, otherwise you get a bottleneck from the OS and you'll get the "low on memory error."
 
Good report of a very relevant test.

I'm sure most gamers think that their FPS increases with a SSD because there is much less (if there are any at all) choppiness that significantly degrade out experience. By using a SSD, the system has one less bottleneck so we can do things a lot faster which gives us the perception of increasing FPS/speed.

All in all, I'm looking forward to my first SSD sometime.
 
[citation][nom]acku[/nom]Ah. I see the confusion. (Pulls out glasses and queue school house rock music....) IOPS * TransferSizeInBytes = BytesPerSecRemember 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...[/citation]

oh so it's bytes!...NOW i get it...thanks a lot! :)

P.S. There's a comment waiting for you here... :hello:
 
Great article. Thanks Andrew.

[citation][nom]Supernova1138[/nom]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.[/citation]
For the sake of proving your claim you could use your spare RAM (I assume you have a ton of it), create a RAM drive, install the game onto it and run.
 
[citation][nom]acku[/nom]IOPS * TransferSizeInBytes = BytesPerSecRemember your units.[/citation]

Wait. How did that make the slightest of difference?
It's like saying a speed of 4000cm/10s=400cm/s but it's not the same as 40m/10s=4m/s.

Since 4KB=4096B, then 50k*4096=204,800,000 B/s = 200,000 KB/s = 195.3125 MB/s.

Which takes us back to square 1.
[citation][nom]ojas[/nom]That's hardly possible, but 4KB transfers at 50k IOPS at 195 MB/s is a fair calculation.[/citation]

Either i'm missing something in my maths or...
 
[citation][nom]K2N hater[/nom]For the sake of proving your claim you could use your spare RAM (I assume you have a ton of it), create a RAM drive, install the game onto it and run.[/citation]
And then install it again after every re-boot?
 
[citation][nom]ojas[/nom]And then install it again after every re-boot?[/citation]
Guess he doesn't restart his PC often. For the case he does... Having a backup/imaging task scheduled to run every time the system has to be restarted along with a restore upon system is up isn't a problem concerning gameplay.
 
[citation][nom]ojas[/nom]Wait. How did that make the slightest of difference?It's like saying a speed of 4000cm/10s=400cm/s but it's not the same as 40m/10s=4m/s.Since 4KB=4096B, then 50k*4096=204,800,000 B/s = 200,000 KB/s = 195.3125 MB/s.Which takes us back to square 1.Either i'm missing something in my maths or...[/citation]

The IOPS you see quoted on spec sheets is only for random transfers. Not for sequential. For a mixed workload, you'd need to calculate the weighed average of the transfer and multiply against the number of read + write operations.
 
i am sorry but, without any intention to troll, your ssd articles are still mostly worthless for the average user.

No average gamer cares about the technical data, we care about results. Log and graph both the load times and the fps over a period of time doing the same activities, both with hd's and with ssd's.

The technical data will not tell us if the ssd provides a more consistent experience, the actual graph of second by second fps will tell us that.
 
Just to prove my point, here is a link to your first article:
http://www.tomshardware.com/reviews/ssd-gaming-performance,2991.html

Now as silly as it is, see the ammount of likes you got, and compare with pretty much any similar article on the videocard section.

The difference isn't only because gpu's in general are "more interesting", it has to do with the way the data is presented. You need to try and reach more people, honestly i'm saying this as constructive criticism.

A section that isn't interesting and consistently gets low hits will equal a section that gets less revenue and eventually may be a section put on the chopping block for good, don't let that happen to you.
 
[citation][nom]Revdarian[/nom]i am sorry but, without any intention to troll, your ssd articles are still mostly worthless for the average user.No average gamer cares about the technical data, we care about results. Log and graph both the load times and the fps over a period of time doing the same activities, both with hd's and with ssd's.The technical data will not tell us if the ssd provides a more consistent experience, the actual graph of second by second fps will tell us that.[/citation]

While I understand your criticism, this article clearly isn't for you. It's not for the technically faint of heart. Second, if you understand the constraints of benchmarking, you'd also understand why fps by second would be an impractical benchmark.

Second, we NEVER choose an article based on the amount of likes. Otherwise, I might as well resort to some of the borderline yellow journalism tactics of some of our news posters. Hate on Apple, Love on Linux and Android, talk about tablets, write short 1 page articles, booth babes, etc...

Third, true SSD enthusiasts like Ao1 at Xtremesystems do like our work. I won't publicly comment on traffic, but I can say this series has definitely NOT been low traffic compared to other articles.

I consider this to be the end of the discussion. I am not interested in a public debate. If other websites want to engage and perpetuate the limited knowledge pool of the "average gamer" that's their business. Tom Hardware is hardly for "average gamers." People here simply know more and are interested in knowing more.

Our readers are above average. If our goal is anything, it's about turning the average gamer into the above average. The idea that we have to keep pounding the same old ground, means that the average gamer never learns. That's never good in technology.
 
@acku Roflol, ok you are perfect, keep doing the articles this way. i have no problem whatsoever understanding them, so /shrugs hahahahahah.

And well qinqin00 don't worry trying to explain anything, it is obvious that with these articles all is perfect, there is no way to make them better and reach farther, and specially, there is also no difference whatsoever of stress between loading the single player campaign that is done in less than 10hours versus actually loading and doing some rounds of multiplayer which will be what oughta last the user at least 7 times longer than the multiplayer.
 
[citation][nom]acku[/nom]The IOPS you see quoted on spec sheets is only for random transfers. Not for sequential. For a mixed workload, you'd need to calculate the weighed average of the transfer and multiply against the number of read + write operations.[/citation]

I understand the weighted average part...which is how i hit the initial confusion over the chart, while trying to find out how much percentage consists of how many bytes 😀 Was only using 4KB for simplicity's sake.

My average was coming to around 32 KB, now you list 53% (for the BF3 launch test) as random transfers, so i wonder if i can assume 16KB to be the average size of random transfers? But that seems wrong for some reason.
Plus some of that would be read, and some would be write, which will have an effect too. I can see the futility of my pursuit... :lol:
 
[citation][nom]acku[/nom]Otherwise, I might as well resort to some of the borderline yellow journalism tactics of some of our news posters. Hate on Apple, Love on Linux and Android, talk about tablets, write short 1 page articles, booth babes, etc...[/citation]
:lol:
 
...and still, I'm not sure what causes the choppiness in BF3 for me. On slightly lowered ultra settings, I get decent framerates (30-40), but for some reason the game lags every couple of seconds before going back to its normal framerate. It's turning my game almost unplayable on these settings, and I'm forced to put especially textures but also other settings much lower than I should be able to run. Is my GPU memory exceeded (HD5850, 1GB), or is my HDD (Samsung HD502HJ ATA) the problem? I don't understand, I have never had so may issues with one game before... my entire computer becomes laggy and unresponsive when running BF3/just having run BF3. Anyone know what the issue might be?
 
[citation][nom]AsTheDeath[/nom]...and still, I'm not sure what causes the choppiness in BF3 for me. On slightly lowered ultra settings, I get decent framerates (30-40), but for some reason the game lags every couple of seconds before going back to its normal framerate. It's turning my game almost unplayable on these settings, and I'm forced to put especially textures but also other settings much lower than I should be able to run. Is my GPU memory exceeded (HD5850, 1GB), or is my HDD (Samsung HD502HJ ATA) the problem? I don't understand, I have never had so may issues with one game before... my entire computer becomes laggy and unresponsive when running BF3/just having run BF3. Anyone know what the issue might be?[/citation]
CPU bottleneck?
Check what's happening to your GPU using GPUz. What resolution are you playing at?

And temperatures too, maybe. Log your temps using Core Temp and GPUz, or check for peak values of temps using RealTemp.
 
[citation][nom]ojas[/nom]CPU bottleneck? Check what's happening to your GPU using GPUz. What resolution are you playing at?And temperatures too, maybe. Log your temps using Core Temp and GPUz, or check for peak values of temps using RealTemp.[/citation]

I have an i5 750, which constantly runs at 100% frequency while playing according to performance monitor (I think that's what it's called, but it's in Dutch for me...), but not all of the four cores are completely used, so I don't think that's the problem (because the CPU graph is consistent, and performance is not). Until now, I didn't have anything to see what my GPU did, but I'll download GPUz and see what it says. I'm playing on 1920x1080.

Thanks for your advice... it's just kind of odd that it seems as if my PC can run it smoothly, but just hiccups for no apparent reason. Actually, the first ~5 minutes I play I don't have any issues at all, but then it starts to hiccup more and more frequently... Perhaps I'll contact Tech support.
 
[citation][nom]ojas[/nom]I understand the weighted average part...which is how i hit the initial confusion over the chart, while trying to find out how much percentage consists of how many bytes Was only using 4KB for simplicity's sake.My average was coming to around 32 KB, now you list 53% (for the BF3 launch test) as random transfers, so i wonder if i can assume 16KB to be the average size of random transfers? But that seems wrong for some reason.Plus some of that would be read, and some would be write, which will have an effect too. I can see the futility of my pursuit...[/citation]

Well, you're totally right. :) It's very difficult to separate the read and write data rates. You could calculate overall IOPS by combining read and write operations and then multiplying by the (weighted) average transfer size.

You bought up a very valid point. I'll check to see if there's a way I can separate the data into more discrete information for articles that come after.

Any games you want to see specifically traced for round three?

Cheers,
Andrew Ku
TomsHardware.com
 
[citation][nom]AsTheDeath[/nom]I have an i5 750, which constantly runs at 100% frequency while playing according to performance monitor (I think that's what it's called, but it's in Dutch for me...), but not all of the four cores are completely used, so I don't think that's the problem (because the CPU graph is consistent, and performance is not). Until now, I didn't have anything to see what my GPU did, but I'll download GPUz and see what it says. I'm playing on 1920x1080.Thanks for your advice... it's just kind of odd that it seems as if my PC can run it smoothly, but just hiccups for no apparent reason. Actually, the first ~5 minutes I play I don't have any issues at all, but then it starts to hiccup more and more frequently... Perhaps I'll contact Tech support.[/citation]

My suspect list:
amount of RAM
storage

I doubt it's graphics related. http://www.tomshardware.com/reviews/battlefield-3-graphics-performance,3063-8.html The 5850 does relatively well. I've seen a limited number of scenarios where increasing RAM helps. But this only applies if you're running with 2 GB. BF3 takes up as much as 2 GB of space, and it performs a memory dump between multiplayer map loading and level loading in campaign mode. Storage could be a potential bottleneck given our results in the article. I wonder if there are background processes running that's inducing the sudden drop off.
 


😀 I consider you a good example of why a gamer here is different from a gamer on some of the other hobby-eqsue sites. You want to learn more and that's what I love. While there certainly people here that just want raw benchmark scores, it's even cooler that there are more readers who want to really understand benchmarks and peel back the layers.
 
Status
Not open for further replies.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom