Without corruption?

Do you mean to tell me, DaveP, that you watched all those goodtime.avi music videos over, and over, and over, and over, and over?

Sad man.

sounds encouraging, but could you preform the same test with a larger file... maybe around 600mb? people are saying a 50mb file could be stored in caches somewhere..

btw, any chance of this driver making it to the prerelease? :)

Is this full speed DMA? Or is it a work-around for the supposed bugs in Artica S?

(by workaround, I mean performance-degrading hack like flushing the CPU cache, as opposed to using Artica S to fix the DMA issues without any slowdown)

Maybe someone could stopwatch the time it takes for 10 runs of MD5 on a CD ISO take to get a rough idea of the bandwidth of the device at the moment?

Great work os4 developers .

No, I just posted the news item Which Im now going to scale to fit.

Dave.

Good things...

_________________
-------------------------------
AmigaOS, the last hope...

Thanks Ben for the informations.

This is very good news for A1 owners!

_________________
ikir

Great news guys

_________________
BACK FOR THE FUTURE

http://www.betatesting.it/backforthefuture

Sam440ep Flex 800 Mhz 1 GB Ram + AmigaOS 4.1 Update 6
AmigaOne XE G3 800 Mhz - 640 MB Ram - Radeon 9200 SE + AmigaOS 4.1 Update 6

It came with the first release of Windows 95!

_________________
Sam

Great news guys Even if the driver is still unoptimized, it still looks promising.

Okay, so this test was using UDMA100 but unoptimised. I'll bet it will still make a nice difference in things under OS4 over the previous PIO driver.

Now, when we get PPC P96 to go with this, OS4 should be running real nice. It already feels fast and responsive, but with these additions, it's going to be pretty sweet.

Slightly OT,

But is that a sneaky use of 'custom' colours in Shell?

Great news!

Any speed advantage over the slow PIO modes is a welcome sight.

Looking forward to the pre-release as always!

_________________
OS4.1 + SAM Flex
RIP my A1XE.. that used to have an appetite for batteries!

a few things...

The speed test was run by jusing NON JIT 68k emulation, right?

And the crc check test was only reading data from the HDD? or?

And caches were disabled??


What about the speed of the bus with smaller test files read from HDD internal cache?

_________________
- KimmoK
// For freedom, for honor, for AMIGA
//
// Thing that I should find more time for: CC64 - 64bit Community Computer?

it's still very slow. i'm curious how much will the software patches slow down the machine. maybe you should start to develop on a microA1?

@Valwit

I see 20-40Mbytes/second speeds reported for Peg1 on ANN in same conditions, tesed with 68k apps..
So, we really can not yet say wether 16Mbytes/sec is really that bad...

_________________
- KimmoK
// For freedom, for honor, for AMIGA
//
// Thing that I should find more time for: CC64 - 64bit Community Computer?

Quote:

software patches

Why patches? Only driver, A1 has no dma problem as i know. Simply the dma approach of Linux is different and bugged (x86). Hyperion is working on the drivers for AOS4 and MAI for Linux.
Again as i know (but i could be wrong) AmigaOne has no dma problem.

_________________
ikir

@ikir

More on that..

I think some people are deliberately trying to misinterpret this driver as a "software patch" equivalent to April1 or 2, and others are getting sucked in. The personal slights posted by one or two kind of revealed their agendas

The truth is the architecture of drivers are radically different between Linux and AOS4 and the AOS4 driver was written from scratch using documentation provided by MAI and is not a "patch" and is just a "driver".

However I can understand the frustration of some, mostly those who predicted that AOS4 would not get a working UDMA driver for AmigaONE.

As for the performance, it is just not optimised yet. Interpreting that as a 68k only statement or a compiler flags statement is incredibly short sighted.

It works, now lets see how they progress it. In the meantime those who were holding back from buying to see how AOS4 handles the DMA issue have one less reason not to buy.

For the more retrenched cynics, maybe in time yet more test results will be published.

AmigaOS doesn't have a filesystem cache like in Linux, so excactly WHERE would this be stored?

_________________
This weeks pet peeve:
Using "voltage" instead of "potential", which leads to inventing new words like "amperage" instead of "current" (I, measured in A) or possible "charge" (amperehours, Ah or Coulomb, C). Sometimes I don't even know what people mean.

@olegil

Oh right... I've used additional cache programs too long.

So, one would need to accidentally type addbuffers 100 000 or something like that.

_________________
- KimmoK
// For freedom, for honor, for AMIGA
//
// Thing that I should find more time for: CC64 - 64bit Community Computer?