Gaming Clarification requested for the Card speed results thread

[Tikker]

Does anyone have a definitive list of which read/write speeds are actually important?

I mean, who cares if a card has 44mbps write speed if all that's required is 5mbps

I see people slagging other people's cards already, and I'm just wondering what the actual requirements for smooth gameplay are?

I'm not going to spend an extra 20-50 bucks for better test speeds if the cheap knockoffs are fast enough

or vice versa, if certain numbers are uber important, it would help justify spending the money on a good card

 

[Tortimer]

I'd like to know this as well.

 

[jos7960]

After all the things i've read about it i got the general idea that you can play any game on any card, but the faster your card is, the faster your loading times will be. Ill keep an eye on this thread for anyone who knows more about this.

 

[solange82200]

I could be wrong, but if you have a slow card I've heard that some of your games could freeze, especially demanding ones like GOW. If someone knows for sure, please clarify

 

[raulpica]

I could be wrong, but if you have a slow card I've heard that some of your games could freeze, especially demanding ones like GOW. If someone knows for sure, please clarify

No, they won't freeze, they would just stutter when in .CSO (compressed) format. Still, you can resolve this by overclocking the PSP's CPU during gameplay.

They will freeze only if you've got a REALLY crappy memstick. But if that's the case, over 90% of things (homebrew, games) won't work

 

[solange82200]

If you have a good stick, could you set the PSP's CPU to something lower than 333mh, in order to save battery life? I have mine at 333, but sometimes I wonder if I even need it that fast...

 

[raulpica]

If you have a good stick, could you set the PSP's CPU to something lower than 333mh, in order to save battery life? I have mine at 333, but sometimes I wonder if I even need it that fast...

Yes, I have mine at 222mhz all the time

If you wonder if you really need it, try the same game with 222mhz, if it's the almost the same speed in-game, you can lower it to 222mhz.

 

[Shyn]

Its not so much the "demanding" games, as much as it is the ones that are constantly reading from the iso. I don't think you'll notice much of a difference between a 16MiB/sec read speed and a 13, but you'll start to notice once you hit like..10.

 

[Alastair]

I think I have a knock-off Pro Duo but only notice slow-downs in the GTA games as compressed .isos - they run flawlessly as .iso.

 

[IBNobody]

If you have a good stick, could you set the PSP's CPU to something lower than 333mh, in order to save battery life? I have mine at 333, but sometimes I wonder if I even need it that fast...

If you don't mind an occasional stutter, here's a nifty feature for you.

CWCheat lets you set your bus speeds in-game. The default is '0', which means it uses whatever the bus was set to before it was loaded up. In most cases, this is 222Mhz.

BUT!

The author of CWCheat added in auto bus speed throttling.

Set your CPU bus speed to '-1', and CWCheat will automatically throttle your CPU and FSB speeds based on the load that the CPU is under.

If the CPU usage rises above 80%, CWCheat increases the CPU/FSP speed by a fixed amount. If CPU usage drops below 40% for a fixed time period, CWCheat lowers the CPU/FSP speed. Best of all, you can actually specify all these settings in the INI file.

Using this, you won't drain your battery as fast when, for example, you are reading dialogs in an RPG.

This is just another reason why you should use (and donate to) CWCheat.

 

[Tikker]

Its not so much the "demanding" games, as much as it is the ones that are constantly reading from the iso. I don't think you'll notice much of a difference between a 16MiB/sec read speed and a 13, but you'll start to notice once you hit like..10.

do you know this for sure? or just speculating?

it'd be nice to have some verifiable answers

 

[jelbo]

Wow, that sounds almost too good to be true. Will try that out for sure. Thanks for the info!

 

[Tikker]

The Performance Index is completely different to the previous version. At first the formula
for calculating works now by weighting factors. So the most important test, which is obviously
the 32kB blocksize reading test, has a higher weighting for the PI than for example the 512B
blocksize writing. Also the Filesystem tests were removed, as they appear to be too heavily
dependent on fragmentation and formatting. blackSPEED now only measures raw-speeds, which is
absolutely fragmentation and formatting independent.
Also some speed optimizations were introduced to max out the read/write-performance, especially
on low blocksizes. This includes setting the CPU/BUS to 333/166, loop optimizations and higher
priority-level of the app.
To distinguish the new Performance Index from the old v1.0 one, the base score is now 100. This
is the score that the original Sony 32MB stick has which came with your psp. So if your stick
for example has an Performance Index of 236 this means that it has 236% the speed of the 32MB
stick (meaning the test run 2.36 times faster than with the original 32 MB stick).




bigger PI number is better, and the most important stat is the 32K read speeds

 

[Tortimer]

Major thanks for that info Tikker.

 

[Tikker]

it would still be nice to know absolute minimum scores for acceptable performance