dslreports logo
 
    All Forums Hot Topics Gallery
spc
Search similar:


uniqs
3243

AthlGrond
Premium Member
join:2002-04-25
Aurora, CO

AthlGrond to fw

Premium Member

to fw

Re: Possible to install BF2 on USB stick?

I just saw this and it seems somewhat related to this thread:

Look, no hard drive: flash only

Samsung's solid-state disk drives...

fw
join:2005-09-18

fw to twizlar

Member

to twizlar
said by twizlar:

Yes I did, i'm currently putting all the information together. I will post shortly.
Hey twizlar, how's it going?
Nagrom Nniuq
join:2002-11-19
Springfield, OR

Nagrom Nniuq

Member

I'm still waiting to see the results from this. Did anything ever happen?

twizlar
I dont think so.
Premium Member
join:2003-12-24
Brantford, ON

twizlar to fw

Premium Member

to fw
Yeah sorry, I tested everything. I've just been really busy with work and havn't had any time to put it together. I'll try to do it this week.

Vamp
5c077
Premium Member
join:2003-01-28
MD

Vamp to fw

Premium Member

to fw
If you are looking for speed get a 15,000 rpm drive to install it to.

BonezX
Basement Dweller
Premium Member
join:2004-04-13
Canada

BonezX

Premium Member

trying to see if it is possible to put it on a solidstate memory stick(flash drive) and still be able to play it, putting it on a 15000RPM(which is the speed of SCSI drives, so it makes your suggestion null) completely defeats the purpose of the experiment.
GunnarDanne
join:2002-12-02
Crown City, OH

GunnarDanne to Loker

Member

to Loker
Look, you're wrong. USB drives are all flash based and flash is slow. Unless you have a DRAM based USB drive, which makes no sense to have and probably doesn't exist, loading times will be slow. USB drives also die after a certain amount of writes. I also doubt the access times are faster. USB drives use up cpu cycles. Harddrives have dedicated controllers. Just because it's a solid state drive doesn't mean it will be faster. Thats like saying an external harddrive connected to a USB port will be faster than a PATA-133 drive because USB=Serial and PATA=Parallel.
OHN
join:2003-02-05
Appleton, WI

4 edits

OHN to fw

Member

to fw
Like many said, the read spead of these drives is greatly diffrent then that of a HDD.

The sustained read speed of good flash drivex is about 19 MP/s.
The read speed of a HDD is on average of 50 MP/s-70 MP/s.

Huge diffrance.

BonezX
Basement Dweller
Premium Member
join:2004-04-13
Canada

1 edit

BonezX to GunnarDanne

Premium Member

to GunnarDanne
wow someone is out of the loop.

samsung allready has working(and faster) harddrives in the laptop size(2.5) based on NAND.
quote:
According to the company's press release, the new drive uses a mere 5 percent of the power a regular hard drive does, while outperforming it by 150 percent. The disk reads data at 57MBps and writes at 32MBps. In addition, it weighs less than half of a regular hard drive and runs cooler and more quietly than hard drives. The flash drive also looks like a standard hard drive to the computer, so no drivers are necessary.
»arstechnica.com/news.ars ··· 939.html

this isen't to test if it will run fast on a flash based drive, it's to see if it is possible to run it in general.

and hate to tell you, USB drives have no seek or spinuptime, which greatly reduces their latencies.

also, usb based external enclosure using a samsung 16gb 2.5 flash drive = better then SATA.
GunnarDanne
join:2002-12-02
Crown City, OH

GunnarDanne

Member

Well, yea, I forgot about those SS drives. They're still kinda slow and way too expensive though. But it does show that flash is improving, but still not fast enough for us to justify switching over to it at the moment.

But... what you said, "usb based external enclosure using a samsung 16gb 2.5 flash drive = better then SATA." is incorrect. USB is much slower than SATA, or even younger forms of PATA. Which is why external SATA is becoming more prevalent.

"
Key benefits of eSATA:

Up to three times faster than contemporary external storage solutions: USB 2.0 and Firewire 400 (IEEE 1394)
"
»en.wikipedia.org/wiki/SATA

If you took two identical drives and used one on SATA and one on USB, the SATA one would be faster. Pretend for a minute that you are right and a USB external enclosure is faster than an internal drive... wouldn't we all be using external drives?

The reason why BF2 will not operate well on a USB drive is that USB bottlenecks the bandwidth too much. That and it would be more CPU intensive and that would slow it down even more.

knightry
I'm Not Fat, I'm Festively Plump
join:2002-05-06
San Jose, CA

knightry to fw

Member

to fw
So twizlar, what's the word? I think a lot of us are curious to see the benchmarks you ran..

twizlar
I dont think so.
Premium Member
join:2003-12-24
Brantford, ON

1 recommendation

twizlar to fw

Premium Member

to fw
Game Startup Times.

Device 1 - AMD 4000+, 4GB RAM, 2X36GB Raptors.
6.0 Seconds

Device 2 - AMD 4000+, 4GB RAM, 120GB Seagate SATA HD 7200RPM.
7.2 Seconds

Device 3 - AMD 4000+, 4GB RAM, 4GB Verbatim USB Flash Drive
16.9 Seconds

Device 4 - AMD 4000+, 4GB RAM, 60GB IOStream USB2.0 2.5"
External Drive.

18.3 Seconds

Removal of movies had a significant effect on load times. For both Device 1 and 2 the load times were just slightly above the normal start up times as the movies had to play for a short time before you can cancel it. However with both external solutions the movies took quite some time to load up, > 30 seconds for both the 4gb and 60gb external drives. Based on pure performance the 4gb flash drive was slightly faster for loading the game, however overall throughput seemed to be lacking compared to the 60gb external hard drive when the 200mb or so of movies were included.

Server Load in Time

For the server load in test, I actually took two measurements because of the different hardware requirements for the different parts of the loading in. The first load in part is when the game is loading all the files from the storage device, and initializing the game. This process takes a serious hit on the CPU and Hard drive. The second part (Verifying Client Data) is purely CPU and RAM based.
Device 1 - AMD 4000+, 4GB RAM, 2X36GB Raptors.
Content Loading
18.0 Seconds

Verifying Client Data
12.5 Seconds

Device 2 - AMD 4000+, 4GB RAM, 120GB Seagate SATA HD 7200RPM.
Content Loading
22.0 Seconds

Verifying Client Data
12.5 Seconds

Device 3 - AMD 4000+, 4GB RAM, 4GB Verbatim USB Flash Drive
Content Loading
42.8 Seconds

Verifying Client Data
13.4 Seconds

Device 4 - AMD 4000+, 4GB RAM, 60GB IOStream USB2.0 2.5"
External Drive.

Content Loading
38.7 Seconds

Verifying Client Data
13.9 Seconds

FPS Tests.

From past experience with BF2, RAM seems to be one of the biggest issues with high FPS during gameplay. However whenever the game needs to load content off of the storage device performance can take a large hit based on throughput and seek times.

Device 1 - AMD 4000+, 4GB RAM, 2X36GB Raptors.
Average FPS
87.6

Device 2 - AMD 4000+, 4GB RAM, 120GB Seagate SATA HD 7200RPM.
Average FPS
87.6

Device 3 - AMD 4000+, 4GB RAM, 4GB Verbatim USB Flash Drive
Average FPS
56.5

Device 4 - AMD 4000+, 4GB RAM, 60GB IOStream USB2.0 2.5"
External Drive.

Average FPS
62.9

The actual framerates for all tests ran about the same ~85 FPS or so. However whenever the game grabbed any content from the storage device, both USB devices really hurt the framerates. I was having trouble staying in the air with a jet or chopper whenever it loaded content, the game would slow down to 15-20 FPS and crawl like that for 10 seconds or so until it was finished loading. Quite unplayable in my opinion. I could see a much faster flash device like an INTERNAL flash based hard drive, like samsung's or one of the RAMDrive devices that hooks into the PCI-E slots. However with an external device like a Thumbstick, or USB Hard drive they aren't quite fast enough and the overall protocol doesn't appear to have the necessary bandwidth to provide enough constant performance for higher end games. In any event YMMV based on your hardware specs, however I don't feel anything else on this test system would be holding back the tests at all.

Common Hardware Used.

AMD64 4000+ @ 2.7ghz
4X1GB Corsair XMS2
Powercolor X1900XTX
Samsung 19" 930B LCD
Creative XFI Gamer Sound Card
Logitech G15 Keyboard
Logitech G5 Mouse
Logitech USB350 Headset
Enermax 650W PSU
Lite-On 16X DVD-RW Drive

If anyone has any other questions or wants any other specific tests, just ask. Thanks.

Jovi
Premium Member
join:2000-02-24
Mount Joy, PA

Jovi to fw

Premium Member

to fw
Thanks for taking the time to do this. Nice reading the results.

fw
join:2005-09-18

fw to twizlar

Member

to twizlar
Thanks!
Nagrom Nniuq
join:2002-11-19
Springfield, OR

Nagrom Nniuq to fw

Member

to fw
Thanks man. Very good info. What resolution and graphics settings are you using just out of curiosity?

twizlar
I dont think so.
Premium Member
join:2003-12-24
Brantford, ON

twizlar to fw

Premium Member

to fw
Everything was run on High @ 1280x1024 4xAA
Nagrom Nniuq
join:2002-11-19
Springfield, OR

2 edits

Nagrom Nniuq to fw

Member

to fw
nvm