That depends on the games engine and how efficient it at utilizing system resources (RAM for one thing). In other words with most up to date games out there, storage performance (hard drive performance) is not issue and if the game engine is that dependent on storage performance then by today’s standards it's pretty much considered unacceptable.
Are SSD's an Improvement?
- Thread starter lewisner
- Start date
That depends on the games engine and how efficient it at utilizing system resources (RAM for one thing). In other words with most up to date games out there, storage performance (hard drive performance) is not issue and if the game engine is that dependent on storage performance then by today’s standards it's pretty much considered unacceptable.
SSD
Obviously, that benchmark should be run on an SSD.
Well, it's a tough job, but somebody's gotta do it. This is an OCZ Agility2. Note the access times. Not having mechanical heads flying around makes a huge difference.
The question, as the previous note suggests, is whether any of this turns into actual increased performance that you can see in Trainz. On my machine it does. Using identical TS12 installations -- one on the OCZ Agility2 and one on a standard WD 5400 RPM hard drive -- and using the N&W Appalachian Coal route (widely considered a challenge for CPUs, hard drives, and GPUs) I see noticeable 'stutter' on the hard drive that isn't there using the SSD. The frame rate difference isn't that high, maybe 4 or 5 fps difference at a particularly tough point in the route, but the second and half-second pauses that characterize this route on the hard drive just go away with the SSD.
Obviously, that benchmark should be run on an SSD.
Well, it's a tough job, but somebody's gotta do it. This is an OCZ Agility2. Note the access times. Not having mechanical heads flying around makes a huge difference.
The question, as the previous note suggests, is whether any of this turns into actual increased performance that you can see in Trainz. On my machine it does. Using identical TS12 installations -- one on the OCZ Agility2 and one on a standard WD 5400 RPM hard drive -- and using the N&W Appalachian Coal route (widely considered a challenge for CPUs, hard drives, and GPUs) I see noticeable 'stutter' on the hard drive that isn't there using the SSD. The frame rate difference isn't that high, maybe 4 or 5 fps difference at a particularly tough point in the route, but the second and half-second pauses that characterize this route on the hard drive just go away with the SSD.
Last edited:
"That depends on the games engine and how efficient it at utilizing system resources"
There's the problem, many game developers pay as little attention to the hard drive bottleneck as the average gamer does, so they don't see the point of designing it to load ALL the scenery into RAM when the game starts. The ones that do make a major difference since the smoothness is not dependent on the slowest component in any system, the hard drive.
Bob, my like, me want. Minimum transfer speed is higher than the max WD transfer speed I have now, average transfer speed is 150% of the WD and 300% of the Maxtor. The seek time is where the real kicker is, even if the disk is defragged regularly (as these were before the tests) it takes time for a standard disk drive to find the files - average 18 milliseconds for the Maxtor, 15.5 for the WD, 0.2 for yours?! Wow. Even if the transfer speed was the same the dramatic decrease in seek time pretty much means the end of "stuttering" - even an old video card has little problem keeping up with the pixels, the main reason for stuttering is because for a few milliseconds in every cycle the video card has nothing to send to the screen, it's waiting for the RAM to send something for it to process, the RAM is waiting for the hard disk to find the next set of files.
Budgeting, looking at the prices;
60gig - $90
120gig - $163
240gig - $309
I think I'd go with the 120 gig with a 500 gig SATA as a secondary drive for storage, utilities and games that don't require blazing speed, run the OS and all the simulators from the solid state drive. However I do it I'm sure I'd get more bang for my buck spending $309 on the 240gig drive than spending the same amount on a new CPU, video card, and RAM which would have the same problem as the older stuff - having to wait for data from the hard disk.
There's the problem, many game developers pay as little attention to the hard drive bottleneck as the average gamer does, so they don't see the point of designing it to load ALL the scenery into RAM when the game starts. The ones that do make a major difference since the smoothness is not dependent on the slowest component in any system, the hard drive.
Bob, my like, me want. Minimum transfer speed is higher than the max WD transfer speed I have now, average transfer speed is 150% of the WD and 300% of the Maxtor. The seek time is where the real kicker is, even if the disk is defragged regularly (as these were before the tests) it takes time for a standard disk drive to find the files - average 18 milliseconds for the Maxtor, 15.5 for the WD, 0.2 for yours?! Wow. Even if the transfer speed was the same the dramatic decrease in seek time pretty much means the end of "stuttering" - even an old video card has little problem keeping up with the pixels, the main reason for stuttering is because for a few milliseconds in every cycle the video card has nothing to send to the screen, it's waiting for the RAM to send something for it to process, the RAM is waiting for the hard disk to find the next set of files.
Budgeting, looking at the prices;
60gig - $90
120gig - $163
240gig - $309
I think I'd go with the 120 gig with a 500 gig SATA as a secondary drive for storage, utilities and games that don't require blazing speed, run the OS and all the simulators from the solid state drive. However I do it I'm sure I'd get more bang for my buck spending $309 on the 240gig drive than spending the same amount on a new CPU, video card, and RAM which would have the same problem as the older stuff - having to wait for data from the hard disk.
God thread with useful information. Thanks Bob, for posting the results!
I can't do that benchmark at the moment because I am away from home at the moment but I'll give it a shot when I get back.I'm doing a revamp of my PC now and have recently got an ASUS Maximus IV Extreme Z mobo and will shortly have an Intel i7 2600K CPU so I will also try it then.Thanks for the replies so far!
Got home today and ran the HD Tune scan with the following results below BUT......I also received two terrifying warnings that the SSD was "CRITICAL at 128 degrees C!!
I thought this was nonsense so got the side panel off and gingerly touched the SSD which wasn't even warm! Anyone have any idea why it said this?
![6116908379_a3c55db3a7.jpg](http://forums.auran.com/trainz/[URL="http://www.flickr.com/photos/pinzac55/6116908379/"][IMG]http://farm7.static.flickr.com/6188/6116908379_a3c55db3a7.jpg)
[/URL]HD Tune SSD Scan by pinzac55, on Flickr[/IMG]
I thought this was nonsense so got the side panel off and gingerly touched the SSD which wasn't even warm! Anyone have any idea why it said this?
Most games are Random Read applications not Sequential Read which is what most benchmark programs such as HDTUNE and HDTACH use to test HDD speed and performance. The mentioned benchmark speed tests posted here are not indicative of what games/sims need from the storage solution.
Here is a Random Access scan of my SSD
HD Tune SSD Scan Random Access by pinzac55, on Flickr
And for comparison a Random Access Scan of my WD 300GB Velociraptor though I freely admit I don't have much idea what these figures mean.
HD Tune Velociraptor Scan Random Access by pinzac55, on Flickr
HD Tune SSD Scan Random Access by pinzac55, on Flickr
And for comparison a Random Access Scan of my WD 300GB Velociraptor though I freely admit I don't have much idea what these figures mean.
HD Tune Velociraptor Scan Random Access by pinzac55, on Flickr
Thanks Lewisner... I know you're going to hate hearing this, but you've actually just convinced me with this post that I was utterly right to return the Sandforce based SSD for an intel one. Despite the Sandforce SATA3 chipset being 2x as fast as the Intel SSD chipset on paper, your results show that it's just that - on paper...
Screenshot_20110905_222833 by nikki_archibald, on Flickr
128 centigrade is about 260 fahrenheit, which should have been hot enough to burn your finger. Only thing I can think of if it was really cool is the monitor was reading it wrong. What do the specs call for that model? Quick giggle search seems to suggest solid state drives SHOULD run cooler "since they have no moving parts", but I would question that since more transistors packed into a smaller space would logically generate more heat, not less.
Nikkia said "Thanks Lewisner... I know you're going to hate hearing this, but you've actually just convinced me with this post that I was utterly right to return the Sandforce based SSD for an intel one. Despite the Sandforce SATA3 chipset being 2x as fast as the Intel SSD chipset on paper, your results show that it's just that - on paper..."
Well Nikkia although I know a bit about computers I don't know a lot! One of the UK PC magazines said the OCZ Vertex 3 was the fastest there was ...... maybe it has something to do with the connection speed to my mobo?
I'll try it again when I have the ASUS Maximus IV installed.
Sniper I really can't understand what the deal is with the 128 degrees thing.If it was that hot I would expect it to have crashed? SSD's are supposed to produce very little heat and no noise, and in fact when I performed the test on the Velociraptor you could actually hear it spinning.
Well Nikkia although I know a bit about computers I don't know a lot! One of the UK PC magazines said the OCZ Vertex 3 was the fastest there was ...... maybe it has something to do with the connection speed to my mobo?
I'll try it again when I have the ASUS Maximus IV installed.
Sniper I really can't understand what the deal is with the 128 degrees thing.If it was that hot I would expect it to have crashed? SSD's are supposed to produce very little heat and no noise, and in fact when I performed the test on the Velociraptor you could actually hear it spinning.
I worked on integrated circuits back when the US Navy first started using them in detection, tracking, and fire control circuits. My experience was analog gave false readings occasionally, digital triggered false alarms frequently. Catapult shot in a S3A Viking one night the fire warning lights for both engines came on during the cat stroke, I wet my pants for the first time since I was potty trained, turned out to be a fault in the detection programming. Technology has improved somewhat since those days, but the nature of the beast being what it is they still screw up on occasion. Primary reason modern fighter planes have three fly-by-wire computers, when one wildly disagrees with the other two they're programmed to outvote it and shut it down.
Saw a few bank clocks last winter, on a day the temp dropped to -35F, one clock displayed the time as 9:74AM and the temperature as 254F, another showed the correct time but reported the temperature as ????. :hehe: And BTW, how many times have you heard a car alarm going off and somebody was actually trying to steal the car? :wave:
What I would do is find a couple other hard drive temperature monitoring programs, see what those say. If two report normal temps and one says it's hot enough to use for a welder, I'd say it was okay.
Saw a few bank clocks last winter, on a day the temp dropped to -35F, one clock displayed the time as 9:74AM and the temperature as 254F, another showed the correct time but reported the temperature as ????. :hehe: And BTW, how many times have you heard a car alarm going off and somebody was actually trying to steal the car? :wave:
What I would do is find a couple other hard drive temperature monitoring programs, see what those say. If two report normal temps and one says it's hot enough to use for a welder, I'd say it was okay.
It gets weirder! I downloaded SpeedFan, ran it and saw 128 degrees and a little "flame" symbol! Touched the HDD and it is as cool as a cucumber. I omitted to mention that when I built my PC I installed an Aerocool V12XT fan controller/temperature monitor, but when I installed the new HDD I didn't attach the sensor since it was a bit fiddly. I've reattached the sensor and it shows the HDD varying between 24 and 26 degrees C.
I've done a bit of Googling and found this useful thread
http://www.wilderssecurity.com/archive/index.php/t-165485.html
with the interesting quote from Ice_Czar "A changing wild reading is generally a misidentification, a static wild reading is generally something trying to be read that isnt there."
This is totally static and even with my limited knowledge of PC's I'd expect it to increase from the time I powered it up?
I've done a bit of Googling and found this useful thread
http://www.wilderssecurity.com/archive/index.php/t-165485.html
with the interesting quote from Ice_Czar "A changing wild reading is generally a misidentification, a static wild reading is generally something trying to be read that isnt there."
This is totally static and even with my limited knowledge of PC's I'd expect it to increase from the time I powered it up?
Last edited:
"a static wild reading is generally something trying to be read that isnt there"
That would be my guess as to the problem with that temp monitor program, instead of displaying an error message that it can't read the hard drive temp, it just displays the maximum temperature it's capable of reading.
128 is suspect for that very reason - 128 is hexidecimal 80, binary 10000000, evenly divisible by 16 and exactly half of 256. If it said 131 degrees or 124 degrees instead of 128 it would be less suspect in my opinion. Again, find some other similar programs and run the tests with those, see what they say.
That would be my guess as to the problem with that temp monitor program, instead of displaying an error message that it can't read the hard drive temp, it just displays the maximum temperature it's capable of reading.
128 is suspect for that very reason - 128 is hexidecimal 80, binary 10000000, evenly divisible by 16 and exactly half of 256. If it said 131 degrees or 124 degrees instead of 128 it would be less suspect in my opinion. Again, find some other similar programs and run the tests with those, see what they say.
OK.....my new machine was booted up this afternoon and the results are promising. Windows Experience gives it 7.7 out of 7.9 and that is without any tweaking or overclocking, just basic settings.
The first two pix are the SSD scan -
HD Tune SSD scan new by pinzac55, on Flickr
Then the SSD Random Access scan -
HD Tune SSD Random Access Scan New by pinzac55, on Flickr
And since I have decided to omit the 300GB Velociraptor out of the new build, a scan of the WD Caviar Black 1TB storage drive -
HD Tune WD Caviar Black Scan New by pinzac55, on Flickr
I think an average of 472Mb/s is OK? :hehe:
The first two pix are the SSD scan -
HD Tune SSD scan new by pinzac55, on Flickr
Then the SSD Random Access scan -
HD Tune SSD Random Access Scan New by pinzac55, on Flickr
And since I have decided to omit the 300GB Velociraptor out of the new build, a scan of the WD Caviar Black 1TB storage drive -
HD Tune WD Caviar Black Scan New by pinzac55, on Flickr
I think an average of 472Mb/s is OK? :hehe:
My new system is well bedded in now and I had a graphic (no pun) demonstration of it's abilities yesterday. My TRS 2009 installation had been copied over from my old Velociraptor HDD to the SSD and although it "ran" it had all kinds of unpredictable behaviour.
So I decided to start afresh and do a clean install of 2009 with of course all the Service Packs.
In the TRS 2009 forum one guy said (I think) that for SP3 you should leave your PC on "all night" to install it ....... it downloaded in just under 3 minutes and installed in TWENTY FIVE minutes!
Running my 2 routes is sublime; there's not a trace of lag or stutter and FPS are around 30-40 FPS even in high detail areas.
So I decided to start afresh and do a clean install of 2009 with of course all the Service Packs.
In the TRS 2009 forum one guy said (I think) that for SP3 you should leave your PC on "all night" to install it ....... it downloaded in just under 3 minutes and installed in TWENTY FIVE minutes!
Running my 2 routes is sublime; there's not a trace of lag or stutter and FPS are around 30-40 FPS even in high detail areas.
The figure 128 suggests to me that the monitor is not getting a valid signal.
128 is a "computer number" like 512 or 64 and is at the end of the 8-bit range, ie 1-128.
EDIT: Sorry, I see Sniper already pointed this out.
I read in a geek forum somewhere that current benchmark programs are not designed to run accurately with SSD's, and that frequent benchmark tests stress SSD's excessively. No idea if it's true or not.
That's my expert opinion:hehe:
I'm going to buy a small SSD, 60GB, and just put Trainz on it and nothing else. It will be a copy of my TRS2010, so if it messes up I can just dump it and go back to my tortoise-like hard drive........
Mick Berg.
Last edited:
johnwhelan
Well-known member
The figure 128 suggests to me that the monitor is not getting a valid signal.
128 is a "computer number" like 512 or 64 and is at the end of the 8-bit range, ie 1-128.
EDIT: Sorry, I see Sniper already pointed this out.
I read in a geek forum somewhere that current benchmark programs are not designed to run accurately with SSD's, and that frequent benchmark tests stress SSD's excessively. No idea if it's true or not.
That's my expert opinion:hehe:
I'm going to buy a small SSD, 60GB, and just put Trainz on it and nothing else. It will be a copy of my TRS2010, so if it messes up I can just dump it and go back to my tortoise-like hard drive........
Mick Berg.
My Trainz folder is currently more than 100 gigs and since you need to rebuild the database sometimes you need more available space for that as well so I suggest 60 gigs might well be on the small side.
Cheerio John
Thanks Mick your idea makes perfect sense! The SSD is still running as sweet as a nut. I would suggest that if you go for an SSD then 120GB would accomodate Windows plus Trainz and quite a bit of other stuff but shop around for the best price. Also make sure you update to the latest Firmware and that you have it connected to SATA 3 cables. There's a guy on YouTube who reviews PC hardware and he admitted that when he got his SSD he wondered why he got crap speeds till he realised he had retained his old SATA 2 cables :hehe: !
The new OCZ REVO drives are interesting and I think that when the price comes down I may need to get one :mop:
The new OCZ REVO drives are interesting and I think that when the price comes down I may need to get one :mop: