@Agafaster

Quote:

got a URL ?

Here.

@KimmoK
Quote:

I'm pretty sure it is not signifficantly faster, if it's faster at all..

In theory there ought be no major advantage to DDR since SDR can saturate the FSB, but the memory controller in the ArticiaS doesn't seem to be getting anything like normal efficiency from the SDR. I timed a little bit of memory-intensive code on both my A1 and a PC of similar specs (P3/600, PC133) the other day, and the P3 was about twice as quick.

The ramspeed benchmark tool thats on OS4Depot tells the same story. The A1 pulls an average of 134MB/sec on memory operations, while the PC is doing 289MB/sec. On the ramspeed web site they have a pile of results, including one for an old 266MHz Pentium II with PC66 memory, which gets 157MB/sec.

So the A1 is performing like it has 66MHz memory, not 133MHz. That leaves a lot of headroom on the FSB for DDR memory to fill.

Quote:

s the memory interfaced to CPU via 133Mhz bus? (I assume so because there is no DDR interface on the 750CXe CPU that Adam uses...)

Should be 133MHz, that's as fast as the CXe is rated. I dare say the chip would work just fine at 166, if the rest of the board supports that.

Quote:

What northbridge is Adam going to use?

Your guess is as good as mine on that one, but the scanty information that's been released suggests it could be one of the Marvell Discovery II family, because the specs (AGP 1x, DDR) are quite similar to the Pegasos 2.

I don't know of any MAI part, real or vapour, which fits with that information.

_________________
Who do you serve, and who do you trust? - Galen

@DrBombcrater



You used just like ramspeed -b 1 ??

Intresting, try to make some comparisons ...


Miikka

@DrBombcrater

Ta. sounds like the AmigaOne1200 / AmiJoe. has he got a website ?

_________________
XH558 - the worlds last flying Vulcan.
ok, its actually XL426 in the picture but you know what I mean.

@DaveP

hi i sent my a1 to my dealer before xmas he's still working at it found rom probs- says if it doesn't work soon he's putting in a new mobo - i'm just paying shipping he's paying for the work and parts. that's the deal he offered when he said i needed the upgrades. hope to hear some good news thos week. i'm using an old a2000 ( bought it the month they first came on sale in Winnipeg) it's old slow and reliable. the a1 was new fast and reliable hopefully it will be that when i get it back.

painter

@DrBombcrater

Quote:

DrBombcrater wrote: @olegil Indeed, but I can't see many other possibilities right now. It could be a driver bug, but that's hardly a brighter prospect because it would depend on some functional difference between the SE and the XE big enough to trigger the bug. That means drivers would need to be specifically tailored to work with the SE. I can't see that happening beyond the default set of OS4 drivers, and not even then if nobody knows exactly what is killing the current drivers.

Here is a theory: I remember well that my reset botton worked pretty well in very early Uboot versions. Back then I was hearing XE users complaining about reset botton not working. Then, Uboot got updated and they were all happy that the botton is working. Guess what?! At the same time my SE reset botton stopped working! It never worked since then, and AmigaOS reboot started crashing!

My point is maybe there is a difference between how Uboot should initialize some things on SE and XE. It's just a thought, I maybe VERY wrong!

Firbodi

_________________
MicroA1 G3, 512MB RAM, DVD-RW; and A1-SE G3,512MB RAM, CD-RW & DVD-ROM
Avatar: Perspolis, Capital of Achaemenid Empire 500 B.C.

@firbodi

I don't hink you're wrong at all



(3:30 am here, I just woke up from having a huge spider removed from my back with some rather big pliers in a nightmare for some reason didn't instantly fall asleep again. I know, Freud would probably have some naughty things to say Gotta be up in about 2 hours, cya)

_________________
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

Sleep tight!

_________________
Site admins are people too..pooff!

@tomazkid

Thanks
I'm up again now, though. Can't say I feel refreshed from a good nights sleep

_________________
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.

@DrBombcrater

What settings in ramspeed are you using as mine are way faster than yours.
800Mhz 750FX A1XE.

AFAIK Adams is not using a Northbridge chip.

Also the coder of ramspeed say's it is not valid to compare across OS's

Kgrach

@bombcrater

After running a few tests I doubt the validity of Ramspeed
my test results on a A1XE 800mhz G3 with single stick of 512MB SDRAM.
For laughs I will try it again with two gigs of SDRAM tomorrow.

4.Workbench:> Data:ramspeed -b 4 -m 1
RAMspeed (UNIX) v2.3.0 by Rhett M. Hollander (Alasir Enterprises), 2002-04

4Gb per pass mode

FL-POINT & WRITING 1 Kb block: 6400.00 Mb/s
FL-POINT & WRITING 2 Kb block: 6400.00 Mb/s
FL-POINT & WRITING 4 Kb block: 5688.89 Mb/s
FL-POINT & WRITING 8 Kb block: 5688.89 Mb/s
FL-POINT & WRITING 16 Kb block: 5688.89 Mb/s
FL-POINT & WRITING 32 Kb block: 5688.89 Mb/s
FL-POINT & WRITING 64 Kb block: 1969.23 Mb/s
FL-POINT & WRITING 128 Kb block: 1969.23 Mb/s
FL-POINT & WRITING 256 Kb block: 1969.23 Mb/s
FL-POINT & WRITING 512 Kb block: 287.64 Mb/s
FL-POINT & WRITING 1024 Kb block: 219.91 Mb/s
4.Workbench:> Data:ramspeed -b 2 -m 1
RAMspeed (UNIX) v2.3.0 by Rhett M. Hollander (Alasir Enterprises), 2002-04

4Gb per pass mode

INTEGER & READING 1 Kb block: 2844.44 Mb/s
INTEGER & READING 2 Kb block: 3011.76 Mb/s
INTEGER & READING 4 Kb block: 3011.76 Mb/s
INTEGER & READING 8 Kb block: 3011.76 Mb/s
INTEGER & READING 16 Kb block: 3011.76 Mb/s
INTEGER & READING 32 Kb block: 3011.76 Mb/s
INTEGER & READING 64 Kb block: 1600.00 Mb/s
INTEGER & READING 128 Kb block: 1600.00 Mb/s
INTEGER & READING 256 Kb block: 1551.52 Mb/s
INTEGER & READING 512 Kb block: 426.67 Mb/s
INTEGER & READING 1024 Kb block: 290.33 Mb/s


I somehow doubt that these numbers are in any shape correct.
As the top numbers exceed the bandwidth of expected on a DUAL DDR 2700 system


Kgrach

@kgrach

>I somehow doubt that these numbers are in any shape correct.
>As the top numbers exceed the bandwidth of expected on a DUAL DDR 2700 system

If it just access L1/L2 cache, these high speeds are ok ??

That is, what I think :)

EDITED: Ok. they are HIGH...

I tested ramspeed -b 1 and I got similar results, than drbombcrater. With G3SE/666MHz.

Bit more discussion is needed, before we judge ArticiaS next to crap :) . It is just software test...


Miikka

Last edited by mlehto on 12-Jan-2005 at 06:37 AM.
Last edited by mlehto on 12-Jan-2005 at 06:26 AM.
Last edited by mlehto on 12-Jan-2005 at 06:23 AM.
Last edited by mlehto on 12-Jan-2005 at 06:08 AM.
Last edited by mlehto on 12-Jan-2005 at 06:06 AM.

@kgrach

My G4 numbers are quite a bit different than yours on G3 (FL/WR from 2179 down to 545, and INT/RD from 3303 down to 209). The larger the block, the less impact of the cache(s) I expect.

Try instead 'ramspeed -b 3 -l 5' and you will probably see consistent numbers about 150-160Mb/sec like DrB. has noted.

Scott Woodall

@kgrach

Quote:

kgrach wrote: @bombcrater After running a few tests I doubt the validity of Ramspeed my test results on a A1XE 800mhz G3 with single stick of 512MB SDRAM. For laughs I will try it again with two gigs of SDRAM tomorrow. 4.Workbench:> Data:ramspeed -b 4 -m 1 RAMspeed (UNIX) v2.3.0 by Rhett M. Hollander (Alasir Enterprises), 2002-04 4Gb per pass mode FL-POINT & WRITING 1 Kb block: 6400.00 Mb/s FL-POINT & WRITING 2 Kb block: 6400.00 Mb/s FL-POINT & WRITING 4 Kb block: 5688.89 Mb/s FL-POINT & WRITING 8 Kb block: 5688.89 Mb/s FL-POINT & WRITING 16 Kb block: 5688.89 Mb/s FL-POINT & WRITING 32 Kb block: 5688.89 Mb/s

These are in L1 data-cache. The size of the L1 data-cache is 32kB.
Quote:

FL-POINT & WRITING 64 Kb block: 1969.23 Mb/s FL-POINT & WRITING 128 Kb block: 1969.23 Mb/s FL-POINT & WRITING 256 Kb block: 1969.23 Mb/s

These are in L2. The size of L2 is 256kB.
Quote:

FL-POINT & WRITING 512 Kb block: 287.64 Mb/s FL-POINT & WRITING 1024 Kb block: 219.91 Mb/s

These are in RAM. The first one is a tad high, though.
Quote:

I somehow doubt that these numbers are in any shape correct. As the top numbers exceed the bandwidth of expected on a DUAL DDR 2700 system

L1 data cache SHOULD bloody well exceed bandwidth of DDR 2700

_________________
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.

@hooligan

Quote:

hooligan wrote: ... Sure.. stricktest set of Quality Assurance certifications in the whole wide world

The German word "Stricktest" would mean "knitting contest" in English...

_________________
Ciao

Dandy
__________________________________________
If someone enjoys marching to military music, then I already despise him.
He got his brain accidently - the bone marrow in his back would have been sufficient for him!
(Albert Einstein)

@swoodall

Thats right:

5-benchmark INTmem LongRun mode

Benchmark #1:
INTEGER Copy: 129.58 Mb/s
INTEGER Scale: 129.42 Mb/s
INTEGER Add: 132.24 Mb/s
INTEGER Triad: 137.79 Mb/s
---
INTEGER AVERAGE: 132.26 Mb/s

__________________________________


5-benchmark FLOATmem LongRun mode

Benchmark #1:
FL-POINT Copy: 142.17 Mb/s
FL-POINT Scale: 141.58 Mb/s
FL-POINT Add: 144.97 Mb/s
FL-POINT Triad: 145.11 Mb/s
---
FL-POINT AVERAGE: 143.46 Mb/s


Anyway results are now more human readable form.

Have to ask, how much code optimization make room for speed improvements ?? If we now measure quality of code instead of quality of bus/chip ?? Just thinking ...

Maybe better to install linux again and test it against linux.

I try to find one UCHI controller and memory stick today... same story, testing with OS4 and Linux.


Miikka

@DaveP

Quote:

DaveP wrote: ... A1-XE and SE's effected were always labelled beta, developer, earlybird with the warning that they might and would contain problems. You know, like the Pegasos 1. ...

I regret to say it - but that wouldn't keep any judge in a country, where a 24 month warranty is prescribed by law, away from sentencing Eyetech, as they officially sold the MoBo's to *END USERS*.

So the law for consumer protection applies.
Sorry, but this is the legal situation - no matter how desperately you are looking for excuses for Eyetech...

_________________
Ciao

Dandy
__________________________________________
If someone enjoys marching to military music, then I already despise him.
He got his brain accidently - the bone marrow in his back would have been sufficient for him!
(Albert Einstein)

@olegil

If this makes some intrest to you. Made with cut'n paste, so values are correct, I think. Anyway same, than in original email, what I got. My first impression was, that they are not correct.


G3SE/666 / uA1/G3/800

INTEGER & WRITING 128 Kb block: 1150.56 / 1484.06 Mb/s
INTEGER & WRITING 256 Kb block: 1113.04 / 1484.06 Mb/s
INTEGER & WRITING 512 Kb block: 121.04 / 1473.38 Mb/s
INTEGER & WRITING 1024 Kb block: 120.61 / 161.64 Mb/s
INTEGER & WRITING 2048 Kb block: 120.61 / 123.37 Mb/s

INTEGER & READING 128 Kb block: 1233.73 / 1587.60 Mb/s
INTEGER & READING 256 Kb block: 1197.66 / 1575.38 Mb/s
INTEGER & READING 512 Kb block: 199.22 / 1587.60 Mb/s
INTEGER & READING 1024 Kb block: 197.49 / 1039.59 Mb/s
INTEGER & READING 2048 Kb block: 197.68 / 199.03 Mb/s



FL-POINT & WRITING 128 Kb block: 1402.74 / 1914.02 Mb/s
FL-POINT & WRITING 256 Kb block: 1356.29 / 1932.08 Mb/s
FL-POINT & WRITING 512 Kb block: 122.93 / 1914.02 Mb/s
FL-POINT & WRITING 1024 Kb block: 122.27 / 1008.87 Mb/s
FL-POINT & WRITING 2048 Kb block: 122.27 / 124.80 Mb/s

FL-POINT & READING 128 Kb block: 1528.36 / 2155.79 Mb/s
FL-POINT & READING 256 Kb block: 1484.06 / 2155.79 Mb/s
FL-POINT & READING 512 Kb block: 209.41 / 2155.79 Mb/s
FL-POINT & READING 1024 Kb block: 207.71 / 1248.78 Mb/s
FL-POINT & READING 2048 Kb block: 207.71 / 199.03 Mb/s


uA1 is not mine, so I don't know about mount of L2 cache. I asked but owner didn't know.


Miikka

Last edited by mlehto on 12-Jan-2005 at 09:31 AM.
Last edited by mlehto on 12-Jan-2005 at 09:14 AM.

@DrBombcrater

Quote:

DrBombcrater wrote: ... The present situation must be frustrating for people with affected A1-XEs, but it's downright scary for us SE owners.

... and really encouraging for potential, future customers.

_________________
Ciao

Dandy
__________________________________________
If someone enjoys marching to military music, then I already despise him.
He got his brain accidently - the bone marrow in his back would have been sufficient for him!
(Albert Einstein)

@DrBombcrater

Quote:

In theory there ought be no major advantage to DDR since SDR can saturate the FSB, but the memory controller in the ArticiaS doesn't seem to be getting anything like normal efficiency from the SDR.

Well usage of DDR in Amiga1 would have no speed advantage but DDR are *MUCH* cheaper than SDR and easier to get...

@olegil
Quote:

huge spider removed from my back with some rather big pliers

What movie have you been looking to have such nightmares?

_________________
A computerworld without MS products and Windows!
Connect your Amigas ...
...The Red ONE-A1XE G4 - A3000T- A3000 - A4000 - A2500- A1000 - A600 - CDTV - CD32...
and your PDAs and laptops ...
Psion 5mx Pro - Psion NetBook - Apple iPhone - MacBook Pro