Disk speed impact on TS2010

J

johnwhelan

Well-known member
Fresh install first on Intel SSD using

Trainz.exe -driver=41078:1001300 -intro=disable -Render=DirectX

32.35 frames per second over five minutes timed with a stopwatch. 1600 by 900 otherwise default settings with only default built in assets on build 41615.

The SSD had Win 7 upgrade installed then TS2010 and has 0% fragmentation.

Fresh install first on Samsung 2 terebyte drive, not reknown for its performance.

Trainz.exe -driver=41078:1001300 -intro=disable -Render=DirectX

31.53 frames per second over five minutes timed with a stopwatch. 1600 by 900 otherwise default settings with only default built in assets on build 41615.

Virgin Samsung disk so no fragmentation.

So yes SSDs are faster but not by much.

Cpu W3520 which is roughly an i7 at 2.66 mhz. 6 gigs of triple channel DDR3 memory, and an ATI 3850 1 gig DDR5 memory.

I have a 300 gig Raptor on the machine so I'll do that next but thought I'd run the two extremes first.

Cheerio John
 
What is the overall system state? Do you run anti-virus? Are you running lots of background processes? Just wondering.

It's surprising that an SSD doesn't fare much better than a good drive. OTOH, it isn't totally surprising because TS2010 never seemed to be much of a disk-muncher to me. I would venture to say, from your experiment, that a fast hard disk plays relatively little role compared with factors such as CPU, video card, appropriate OS, streamlining, etc.

It would be nice if we could nail down exactly what 2010 likes and doesn't and what works and what doesn't, and present a tweaked set of specifications accordingly.
 
How about when OFE/Commit, doing an EDR or when it's loading a route/session?

I would have thought the SSD would be significantly better there.
 
RRSignal said:
What is the overall system state? Do you run anti-virus? Are you running lots of background processes? Just wondering.

It's surprising that an SSD doesn't fare much better than a good drive. OTOH, it isn't totally surprising because TS2010 never seemed to be much of a disk-muncher to me. I would venture to say, from your experiment, that a fast hard disk plays relatively little role compared with factors such as CPU, video card, appropriate OS, streamlining, etc.

It would be nice if we could nail down exactly what 2010 likes and doesn't and what works and what doesn't, and present a tweaked set of specifications accordingly.
Click to expand...

It's a brand new system Win 7 64 bit home pre and I used a session that's built in so others can compare their systems directly. Background is basically APC power monitor and that's it. I think I had Microsoft antivirus disabled but I'll dup the Raptor test with and without to see what the impact is.

Cheerio John
 
johnwhelan said:
Fresh install first on Intel SSD using

Trainz.exe -driver=41078:1001300 -intro=disable -Render=DirectX

32.35 frames per second over five minutes timed with a stopwatch. 1600 by 900 otherwise default settings with only default built in assets on build 41615.

The SSD had Win 7 upgrade installed then TS2010 and has 0% fragmentation.

Fresh install first on Samsung 2 terebyte drive, not reknown for its performance.

Trainz.exe -driver=41078:1001300 -intro=disable -Render=DirectX

31.53 frames per second over five minutes timed with a stopwatch. 1600 by 900 otherwise default settings with only default built in assets on build 41615.

Virgin Samsung disk so no fragmentation.

So yes SSDs are faster but not by much.

Cpu W3520 which is roughly an i7 at 2.66 mhz. 6 gigs of triple channel DDR3 memory, and an ATI 3850 1 gig DDR5 memory.

I have a 300 gig Raptor on the machine so I'll do that next but thought I'd run the two extremes first.

Cheerio John
Click to expand...


Nothing new, SSD's performance is heavily based on the controller you connect it to.


If you had defraged the mechanical hard drive with a decent defrag program (O&O Defrag) I wouldn't have been surprised if your results were a wash.


Unless you are on a dedicated PCIe RAID card you are currently not going to see a big improvement. This will change soon when we start to see SATAIII SSD's along with 6Gbps controllers.
 
djt said:
Nothing new, SSD's performance is heavily based on the controller you connect it to.


If you had defraged the mechanical hard drive with a decent defrag program (O&O Defrag) I wouldn't have been surprised if your results were a wash.


Unless you are on a dedicated PCIe RAID card you are currently not going to see a big improvement. This will change soon when we start to see SATAIII SSD's along with 6Gbps controllers.
Click to expand...

Intel X25 SSD, normally recognised as having a reasonable controller. The mechanical drives were brand new, so completely empty therefore there was no need to defrag them.

Cheerio John
 
johnwhelan said:
Intel X25 SSD, normally recognised as having a reasonable controller.
Cheerio John
Click to expand...


I'm talking about the controller on the motherboard not the drive itself.


“Testing” storage performance with a stopwatch and measuring frames per second with a game isn't a “test” of storage solution performance at all to begin with. You are going to need utility that tests “random read/write” not “sequential read” performance
of the storage solution to do a proper test.



The mechanical drives were brand new, so completely empty therefore there was no need to defrag them.
Click to expand...
It doesn't matter if the drives were “brand new” or not, for optimal performance they should be defraged with a decent defrag utility right after any data is installed on them.


You're best bet with those two drives using the motherboard controller is to use the SSD for just the OS and use the Samsung for storage.
 
You must log in or register to reply here.

Similar threads

W
Trainzoptions.txt modifications/additions?
2
Replies
26
Views
30K
StorkNest
S
Back
Top