03-Dec-2022 12:22 GMT.
UNDER CONSTRUCTION
Anonymous, there are 427 items in your selection (but only 177 shown due to limitation) [1 - 50] [51 - 100] [101 - 150] [151 - 200] [201 - 250] [251 - 300] [301 - 350] [351 - 400] [401 - 427]
[Events] Notes about my presentation at the Alchimie 4 showANN.lu
Posted on 29-Sep-2004 10:40 GMT by Stéphane Guillard427 comments
View flat
View list
The html pages I made available was only a starter slideshow for my presentation. Here are a few words about what I shown, and what I said. Hi Gentlemen,

The html pages which are available at my site are just an introduction slideshow that I presented as a starter, it does not relate what I shown and said afterwards. Please find below a couple of notes about my presentation at Alchimie show.

Here is a quick summary of the presentation I did :

- I booted with boot delays set to 1 second for UBoot and for SLB (second level booter), thus less than 20 seconds after power on, we were with a fully loaded Workbench with Amidock.

- I started with the small html slideshow, presented on IBrowse. You can find those pages at the URL above. IBrowse loaded in 2 seconds with its About: page fully displayed. Browsing through the pages of the slideshow was instantaneous.

- While we were at it, I browsed the OS4 install guide, also as fast as can be, must say I also find this responsiveness impressive myself :)

Then I demoed as many things as I had time to during the 2 hours I had. Everything worked, fast & stable, and was smooth and impressed. I showed mainly:

- All os4 system, tools, utilities, prefs & stuff

- MUI, IBrowse, Amitradecenter

- Yam, SimpleMail

- ClassAction (M. Elsner's file manager)

- MakeCD

- MooVid running a DivX

- DVPlayer running an mpeg2

- AudioEvolution 4 with the demo project, cursors auto moving smoothly, the playback was smooth also, with mostly no CPU usage.

- AmiPDF with the AE4 manual and another heavy PDF file, very fast

- USB. I plugged a Microsoft IntelliMouse Optical into my USB hub, and we had 2 mice to control the system

- Crisot's slach5 winning demo (got applauses which it deserved)

- chip's rayrace realtime raytracer demo. When the juggler appeared, audience was impressed, but really took measure of what they were watching when I moved the mouse. Wait for the Altivec version !

- FPSE, AmiDog's PS1 emulator, with an oldskool game which ran nicely ('Cotton')

- some other visual toys I had at hand

The demos only grimreaped twice, and I anticipated the grims before they popped up :

- One was native glsokoban / w3d, when I launch it does a base page access (a “null pointer” bug in glsokoban). I didactically shown the disassembly which is available in the grimreaper window, was a store to r4, r4 was null etc. I clicked on continue, and it all went fine & fast.

- One was frying pan 0.3.1, I shown the app, and at one point I said 'now it should grim’ and it did. It still loaded fine though. I quit the app, clicked on reboot and less than 4 secs after, wb was up with amidock. That was the only reboot of the show.

I forgot to show (because of short time):

- Petunia... Almos, sorry, I had prepared something for that (side by side windowed jit & nonjit runs of voxelspace), but i both forgot, and was asked to stop at this point by the party organizers cause it was already 5:30 pm while I was scheduled until 4pm.

- ArtEffect

- USB with MassStorage (ie USB key or digital camera)

At the end, I had many fair questions to which I answered; my feeling is that the audience really appreciated the effort behind what I shown, and was conscious that we are not far away from a releasable 4.0.

Then came the expected question, 'and why doesn’t DMA work ?'

I said 'All what you saw was DMA since the 1st boot'.

I copied a few 100 MB files in a snap, with zero CPU (thanks to Pete Gordon for the clock/CPU docky, helps a lot). Then I switched to PIO, they saw 4 x slower and 80% CPU.

The audience understood that it was indeed DMA, and that was fast, and that was part of the overall smoothness of what I shown.

Then I explained the things below (this is the reference for my statements, please don’t rephrase or extrapolate or invent or whatever):

- IDE UDMA works on VIA and Articia on AmigaOne SE / XE / µA1 MK2 (as I demoed) ...

- ... except when the Ethernet chip goes online and is used.

- the Ethernet chip only triggers the problem, but it is not at all related to it (a test using a PCI Ethernet shows the same behaviour)

- We have made a driver for a Silicon Image 680 PCI IDE UDMA133 controller chip, this does UDMA 133 nicely, including when Ethernet is used at full speed.

- The fact that a PCI IDE controller solution works, shows that the problem is *not* related to Articia, since PCI DMA is *also* handled by the Articia, and that works.

- The full Alchimie show demo was done using UDMA, both from the VIA and from the Si680, without problem (but with Ethernet off, would the Ethernet have been turned online, I would have had to revert the VIA into PIO before).

- Things are currently under more investigation

In the meantime there are 2 options for existing A1 board owners:

- Use the VIA IDE controller in PIO mode when using Ethernet, and UDMA at other times,

- Purchase a faster (UDMA133) Silicon Image 0680 IDE PCI card (from around $20). This is my personally recommended option as the delivered speed is noticeable faster than the on-board VIA controller in UDMA mode.

Notes about my presentation at the Alchimie 4 show : Comment 251 of 427ANN.lu
Posted by AmiTroll on 30-Sep-2004 10:52 GMT
In reply to Comment 206 (JKD):
@JKD

Excellent man. You nailed it. I said the same way back then and people shouted me down. Even got a post rm'd because i was 'just trolling'. Truth is, these two companys are _small_ and can't do as much as Dell could. I don't know why people expected no problems. And i still think April 2 is not a fix that does the job, and the new A1 bug needs to be fixed as well. If its a driver or hardware, dosen't matter. Both companys owe it to the users, period. Anything else is just bullshit from the same handfull of fools. :)
Notes about my presentation at the Alchimie 4 show : Comment 252 of 427ANN.lu
Posted by priest on 30-Sep-2004 11:03 GMT
In reply to Comment 233 (itix):
Where is that information from?

FSB is not memory bus.

And I think the newer ArtisiaS had improvement for memory compatibility.
Notes about my presentation at the Alchimie 4 show : Comment 253 of 427ANN.lu
Posted by Anonymous on 30-Sep-2004 11:07 GMT
In reply to Comment 250 (itix):
If you are allergic to laughter, then just don't make people laugh by comments like that. ;-)
Notes about my presentation at the Alchimie 4 show : Comment 254 of 427ANN.lu
Posted by priest on 30-Sep-2004 11:07 GMT
In reply to Comment 233 (itix):
As a matter of fact I do not hawe a clue what you meant with that.


Other than that: It would be nice to know why exactly G4 cards are not compatible with Peg1. Is it just because Genesi rather sells Peg2 ? (G4 card FSB clock is locked to 133Mhz ?)
Notes about my presentation at the Alchimie 4 show : Comment 255 of 427ANN.lu
Posted by Mr_Bumpy on 30-Sep-2004 11:22 GMT
In reply to Comment 241 (JoannaK):
Interesting point. Being a PC user for a long time, when I first read about Eyetech's hardware plans and saw the VIA 686B southbridge listed, I thought, "Oh crap! How long before we will be hearing about data corruption problems that only show up when certain other hardware is used at the same time...". I thought this long before the first AmigaOne boards were ever released, and my fears were based solely on my knowledge of the 686B bugs in AMD motherboards. Not to mention that the VIA IDE controller offers rather lousy performance, even when compared to the cheaper, competing ALi southbridge. This and many, many other problems (most of which I've experienced firsthand) have scared me permanently away from VIA chipsets, and I have vowed never to own or recommend a system with the word VIA *anywhere* on or in it. I am not the only one, either. Friends don't let friends buy VIA.
Notes about my presentation at the Alchimie 4 show : Comment 256 of 427ANN.lu
Posted by XraalE on 30-Sep-2004 12:02 GMT
In reply to Comment 215 (MikeB):
@MikeB

> I also believe Wrongpla.net's webmasters are likely just being used by
> bbrv, one of Wrongpla.net most vivid contributors. IMO bbrv is a pretty
> good smooth talker, seemingly lacking generally shared morals. (He
> thinks it's OK to release confidential private information to suit his
> own agenda, speaking for rival companies, pushing people to complain on
> his behalf, etc) I believe some people like KennyR may just be acting
> like his puppets, not really knowing what they are being dragged into.

Where's my bullshit meter? I can hear its alarm going off from inside a cupboard. Overload, overload...!

bbrv contribute to WP? Are you kidding? He contributed ONCE, and only because we asked him to after the Amiga.org thread we found so funny shortly after WP was begun. Please stop writing about things you don't know a goddamn thing about. We have no wish to be pulled into your conspirator fantasy world of how Buck is the evil and rules the planet. Buck doesn't own WrongPlanet, WE do, and WE write when we wish about who we wish. Got that?
Notes about my presentation at the Alchimie 4 show : Comment 257 of 427ANN.lu
Posted by XraalE on 30-Sep-2004 12:05 GMT
In reply to Comment 256 (XraalE):
And speaking of puppets Mike, when is the last time you actually directly talked to any member of KMOS or Amiga Inc you seem heart set on defending with so little that remains of your reputation?
Notes about my presentation at the Alchimie 4 show : Comment 258 of 427ANN.lu
Posted by itix on 30-Sep-2004 12:20 GMT
In reply to Comment 252 (priest):
Some DIMMs are working reliably with Articia only when FSB is clocked down to 100MHz. That is what I meant... the point is that why so many registered SDRAMs do not work on AmigaOne? They work on every other non-Articia machine.
Notes about my presentation at the Alchimie 4 show : Comment 259 of 427ANN.lu
Posted by itix on 30-Sep-2004 12:27 GMT
In reply to Comment 254 (priest):
AFAIK you cant use G3 cards from Peg2 either. I'm not HW guy (yet, studying computer electronics in school) but it has something to do with different FSB. Other matter are G4 cards specifically done for Peg1... Genesi said you must have a matching pair.
Notes about my presentation at the Alchimie 4 show : Comment 260 of 427ANN.lu
Posted by itix on 30-Sep-2004 12:28 GMT
In reply to Comment 253 (Anonymous):
I just realized I'm communicating with 5 years old school boy.
Notes about my presentation at the Alchimie 4 show : Comment 261 of 427ANN.lu
Posted by itix on 30-Sep-2004 12:31 GMT
In reply to Comment 256 (XraalE):
Give it rest... Mike Bouma lives in his own fantasy world. He already trashed this thread, no need to feed this troll anymore.
Notes about my presentation at the Alchimie 4 show : Comment 262 of 427ANN.lu
Posted by MikeB on 30-Sep-2004 12:34 GMT
In reply to Comment 256 (XraalE):
OK, I agree my opinion could have been viewed as a shot in the dark.

So (officially) Bill Buck only once had the opportunity to write for Wrongpla.net as a Columnist. However I still believe he has contributed alot to Wrongpla.net as one of its main advocates. When he pulled away from some main Amiga forums like ANN and Amiga.org he did state he would try to contribute more to Wrongpla.net instead. But OK, it's often hard to know when to take him seriously and when not...
Notes about my presentation at the Alchimie 4 show : Comment 263 of 427ANN.lu
Posted by JoannaK on 30-Sep-2004 12:36 GMT
In reply to Comment 260 (itix):
Don't say that aloud... he might bite you in ankle.. :)
Notes about my presentation at the Alchimie 4 show : Comment 264 of 427ANN.lu
Posted by JoannaK on 30-Sep-2004 12:38 GMT
In reply to Comment 255 (Mr_Bumpy):
Well.. sad to say, but Pegasos has Via Sb-chip too... Luckily it's not the same one (a lot newer chip) and seems to be working quite well..
Notes about my presentation at the Alchimie 4 show : Comment 265 of 427ANN.lu
Posted by Joe on 30-Sep-2004 12:44 GMT
Those Teron boards have had a bad reputation before the Amigaone name. Alan the shopkeeper decided that buying expensive well known buggy hardware was a good move.

For such a high price you would expect something decent and this has lost Eyetech/Hyperion sales.
Notes about my presentation at the Alchimie 4 show : Comment 266 of 427ANN.lu
Posted by Georg Steger on 30-Sep-2004 13:01 GMT
In reply to Comment 219 (Stéphane Guillard):
> Now, a word on the read transfers. For reads, no cache flush is needed at all, neither with hardware snooping nor with software cache handling. Why flush cache to ram when you will be overwriting ram with data read from the device ?

> What is needed for read is cache invalidation. That is, telling the CPU that he has to go to real ram to fetch data, because its cache contains invalid data. And as Bernie Meyer pointed out earlier, this cache invalidation has to be done *before* the read and not after, contrary to intuition. Why ? If for whatever reason, the CPU wants to cache other areas, completely unrelated to the DMA read being done, it might decide to flush cached areas of the read buffer, in order to get free cache lines. If this flush, which happens 'in your back', happens after the DMA controller wrote data to ram, you end up with trash. Simple way to avoid that : make sure that no area of your read buffer is cached, ie invalidate the cachelines that might cover your buffer. This was one bug of mine a long time ago. You see, I'm not even reluctant to admit my own bugs.


I don't know much about this stuff, so it might be a stupid question, but I'm curious: after the cache invalidation (StartDMA call?), what stops the cpu from caching the area again (memory accesses by other tasks?) anytime after that (StartDMA call?), bringing you back to the initial problem.
Notes about my presentation at the Alchimie 4 show : Comment 267 of 427ANN.lu
Posted by Anonymous on 30-Sep-2004 13:10 GMT
In reply to Comment 266 (Georg Steger):
Simple: mark corresponding pages uncacheable. And due to this particular reason implementing a working DMA in Linux is PITA.
Notes about my presentation at the Alchimie 4 show : Comment 268 of 427ANN.lu
Posted by MikeB on 30-Sep-2004 13:13 GMT
In reply to Comment 261 (itix):
I haven't dragged myself into this thread, actually this was done by you and Mr Decanini. Just ignore me in the future and I will happily ignore you.

And IMO your general contribution to this thread wasn't that constructive.

Example:

Article: "the Ethernet chip only triggers the problem, but it is not at all related to it (a test using a PCI Ethernet shows the same behaviour)"

Your reply to the article: " Why? Buggy ethernet driver/chip?"

So, IMO already within comment 3 you started to contribute noise to this thread. Later you even drag the ArticiaS and myself into the discussion creating even more distortion.
Notes about my presentation at the Alchimie 4 show : Comment 269 of 427ANN.lu
Posted by minator on 30-Sep-2004 13:27 GMT
I posted this on the new PPC thread thought it might be relevant here:

Just think, if they decide to use the 8641 / 8641D there'll be no more arguments about the North bridge...

I think they should bite the bullet and actually work together on this, I'm sure a modification could be done for "certification" or whatever and deals struck so end-user pricing is similar.

By using a single board there wouldn't be any arguments about hardware (OS advocates can argue as much as they want) and sharing costs means prices will be lower for all concerned - and that ultimately is what this market needs more than anything else.
Notes about my presentation at the Alchimie 4 show : Comment 270 of 427ANN.lu
Posted by Anonymous on 30-Sep-2004 13:28 GMT
In reply to Comment 260 (itix):
Wow, I must be a real talent then....a 5 years old kid with all that knowledge....
Ah well, maybe i'm just the Einstein of this age.....


;-)
Notes about my presentation at the Alchimie 4 show : Comment 271 of 427ANN.lu
Posted by Anonymous on 30-Sep-2004 13:29 GMT
In reply to Comment 263 (JoannaK):
I'll take the 5 years old story and you comment as a classic example of: "running out of arguments..."
Notes about my presentation at the Alchimie 4 show : Comment 272 of 427ANN.lu
Posted by JKD on 30-Sep-2004 13:32 GMT
In reply to Comment 214 (myself):
You missed a point.....everyone with a broken Peg1 (pre April) got the chance to get it fixed *for free*...no-one had to wait for the Peg 2.
Notes about my presentation at the Alchimie 4 show : Comment 273 of 427ANN.lu
Posted by Christophe Decanini on 30-Sep-2004 13:32 GMT
In reply to Comment 268 (MikeB):
I didn't drag you in the thread. I denied that you were involved in a comment and said that you post anonymously here (which is true) and MAY be in the thread (which was not the case, hence the "may"). So please stop the bullshit Mr Mike Bouma.
Next time I will just not reply when someone brings you in. I guess that then you will accuse me of biased moderation.
Notes about my presentation at the Alchimie 4 show : Comment 274 of 427ANN.lu
Posted by Anonymous on 30-Sep-2004 13:33 GMT
In reply to Comment 264 (JoannaK):
depends of what you call "a lot". The Pegasos southbridge is the notebook version of the 686B. Much parts in the Pegasos southbridge even identfy themself as 686B (PCIinfo)
Notes about my presentation at the Alchimie 4 show : Comment 275 of 427ANN.lu
Posted by Stéphane Guillard on 30-Sep-2004 13:38 GMT
In reply to Comment 266 (Georg Steger):
Georg,

It is very simple : you don't access the buffer while dma'ing to it. It would make no sense, and it is like that on all os'es.

Kind regards,
--
Stéphane
Notes about my presentation at the Alchimie 4 show : Comment 276 of 427ANN.lu
Posted by MikeB on 30-Sep-2004 13:50 GMT
In reply to Comment 273 (Christophe Decanini):
You gave some one-sided story about: "M.Bouma was asked not to post on this site anymore and he did several times using anonymous. He was easy to trace as he was not using proxies." and you suggested I may be one of the anonymous people, despite you know my IP address.

Just stating: "No that wasn't him." [The End] would IMO have been a far more constructive approach. And with regard to bias, I believe everyone knows that you are biased, just as much as everyone knows I am (both due to our own personal experiences and history).
Notes about my presentation at the Alchimie 4 show : Comment 277 of 427ANN.lu
Posted by itix on 30-Sep-2004 13:58 GMT
In reply to Comment 276 (MikeB):
It was me who dragged Mike Bouma to this discussion. I connected your name to the anon poster. My fault. But hey, we can still hit the record.
Notes about my presentation at the Alchimie 4 show : Comment 278 of 427ANN.lu
Posted by Georg Steger on 30-Sep-2004 14:34 GMT
In reply to Comment 275 (Stéphane Guillard):
> It is very simple : you don't access the buffer while dma'ing to it. It would make no sense, and it is like that on all os'es.

Okay, but doesn't caching work on a page basis? And can't different allocations (in OS4) of different apps end up using the same page (also because IIRC of the rater big page size of 64K?) And while porting the GPL SBLive/Audigy AHI driver to a certain other OS a few weeks ago, I noticed that the OS4-specific version of linuxsupport.c/AllocPages() in there does a simple AllocVec(..., MEMF_PUBLIC).

And so if two apps' allocated memory could be in the same page, then "don't access the buffer while dma'ing" might be done by one task (to whom the dma buffer "belongs"), but you don't know what that other task might or might not
do with it's allocated memory at the same time?
Notes about my presentation at the Alchimie 4 show : Comment 279 of 427ANN.lu
Posted by Don Cox on 30-Sep-2004 14:36 GMT
In reply to Comment 231 (Anonymous):
"Instead they invented another April version.
Probably to add hardware coherency, a thing they DID understand.
But they weren't able to track their FSB and G4 problems."

It's easy to sit in an armchair and make rude remarks about MAI's or BPlan's board design skills. The fact is that designing a modern motherboard is extremely difficult, especially on a limited budget when you can't afford to have large numbers of prototypes xmade up and tested.

It is not surprising that early production batches (which are all there are) of AmigaOnes and Pegasoses have some bugs.
Notes about my presentation at the Alchimie 4 show : Comment 280 of 427ANN.lu
Posted by Don Cox on 30-Sep-2004 14:41 GMT
In reply to Comment 247 (MikeB):
"Yes, but for me OS advancements are more important than backwards compatibility!"

Not for me. I want to run programs and actually use the computer, not sit around admiring the operating system and fiddling with the colours on the Workbench.
Notes about my presentation at the Alchimie 4 show : Comment 281 of 427ANN.lu
Posted by Don Cox on 30-Sep-2004 14:42 GMT
In reply to Comment 247 (MikeB):
"The advancement from version 1.3 to 2.0 already broke alot of compatibility."

Give examples of programs that worked in 1.3 and not in 2.0.
Notes about my presentation at the Alchimie 4 show : Comment 282 of 427ANN.lu
Posted by JoannaK on 30-Sep-2004 15:04 GMT
In reply to Comment 271 (Anonymous):
Well.. I had to point out the obvious cause some people are simply too stoopid to see it wihtout help.
Notes about my presentation at the Alchimie 4 show : Comment 283 of 427ANN.lu
Posted by JoannaK on 30-Sep-2004 15:10 GMT
In reply to Comment 274 (Anonymous):
Well.. This VIA chip does work quite well on Pegasos.. AC97, Ide-DMA (with nice speed), USB. It does not mess with On.board ethernet etc... You tell me. Are Mos and Pegasos-linux coders so much better or what makes all the difference? Is it hardware, software, luck, skill ... Obviously something must be there.. You tell me ..
Notes about my presentation at the Alchimie 4 show : Comment 284 of 427ANN.lu
Posted by itix on 30-Sep-2004 15:12 GMT
In reply to Comment 281 (Don Cox):
Actually it was not the OS dropping compatibility... software was just coded wrong. Even simple Kickstart 1.2 -> 1.3 upgrade broke many demos and games. Another thing is if you drop obsolete technology like BCPL support. It died long long ago and is no longer needed.
Notes about my presentation at the Alchimie 4 show : Comment 285 of 427ANN.lu
Posted by Agima on 30-Sep-2004 15:28 GMT
In reply to Comment 281 (Don Cox):
>>Give examples of programs that worked in 1.3 and not in 2.0.

Sure, you're right there are none. Huh, I wonder why the dual kick start boards that let you have 1.3 and 2.0 were so popular if there was no software that didn't work under 2.0 that did work on 1.3 machines then. I can already hear you argument and I'm sure you're right, it was because of all the 'Kickstart Banging' games... yeah, that's it.
Notes about my presentation at the Alchimie 4 show : Comment 286 of 427ANN.lu
Posted by MikeB on 30-Sep-2004 15:44 GMT
In reply to Comment 281 (Don Cox):
There were loads of programs that didn't work anymore I agree that this was mostly due to the fault of bad programming and that's part of the reason why programs like Relokick, TUDE and Degrader were developed. Also today especially loads of games/demos still require such programs to be able to run them properly on later systems (BTW WHDLoad rules!!).

Of most professional software newer more compatible versions were created (there were issues with some word processors, spreadsheets, movie players, etc) and that's what I hope to see happening with AmigaOS4 as well. For nearly full backwards compatibility I would rather use the AmigaOS4 version of UAE.

I am sure you can still find some software on Aminet that hasn't been updated yet since the 80s. (although especially incompatible ones are more likely to have been updated) Give them a try probably you will still find a few which won't work properly with your current configuration (without the above mentioned tricks).

Anyway amongst the reasons why features like Virtual Memory and Resource Tracking were never implemented in classic AmigaOS releases in the past was because of software compatibility issues this would cause. However it's now more than 12 years since the kernel functionality has changed (Kickstart 3.1) and IMO today we can't do without these features within any modern operating system.
Notes about my presentation at the Alchimie 4 show : Comment 287 of 427ANN.lu
Posted by Christophe Decanini on 30-Sep-2004 16:00 GMT
In reply to Comment 276 (MikeB):
I stated facts, you came with lame excuses and telling I was trolling.
My bias is that I don't like liars and hypocrites like you. It as nothing to do with your "color".
I have been criticized for my moderation equaly from both sides which should show how biased I am in my moderation.
As for my personal opinions, I am as free as any user here to state them and if any other moderator think that I go over the lines and against this site policies they are free to moderate me.
Notes about my presentation at the Alchimie 4 show : Comment 288 of 427ANN.lu
Posted by Agima on 30-Sep-2004 16:01 GMT
In reply to Comment 281 (Don Cox):
>>Give examples of programs that worked in 1.3 and not in 2.0.

How about a list of softare that works on XP SP1 but does not work with XP SP2?

-Command Antivirus
-Encyclopedia Britannica 2000 Deluxe
-eTrust EZ Armor
-Live Motion (from Adobe)
-MPEGcraft DVD
-NOD32 for Microsoft Windows
-PaperPort Deluxe
-Virtual PC (runs but unusably slowly copmpared to XP SP1)
-several version of ZoneAlarm
-NBA LIVE 2000
-BitDefender
... and on and on....

I guess Windows XP SP2 is not "REALLY" the Windows OS. So sad :(
Notes about my presentation at the Alchimie 4 show : Comment 289 of 427ANN.lu
Posted by RedOS rulez !!! on 30-Sep-2004 16:10 GMT
In reply to Comment 219 (Stéphane Guillard):
"the µA1 'C' alias MK3 (and further machines) has proper snoop signaling implemented"

Wow! Eyetech Ltd reverse engineered an April fix? I'm impressed. Good for you. The April fix for an MicroAmigaOne is an essential.

1 point for the RedOS !!! Yuhuuu!! Long live AmigaOne with the April fix!
Notes about my presentation at the Alchimie 4 show : Comment 290 of 427ANN.lu
Posted by Fabio Alemagna on 30-Sep-2004 16:19 GMT
In reply to Comment 289 (RedOS rulez !!!):
Indeed, I'm amazed not many noticed that sentence.
Notes about my presentation at the Alchimie 4 show : Comment 291 of 427ANN.lu
Posted by Christophe Decanini on 30-Sep-2004 16:34 GMT
In reply to Comment 290 (Fabio Alemagna):
I did notice it but I guess talking about it will be considered by certain as "trolling".
The articia has no problem but does not have proper snoop signaling ?
As for terrasoft dropping the Teron it was certainly not because the articia had a problem and could not be supported in Linux ...
Everyone can understand why Genesi dropped the articia and said it was buggy and everyone understand that hyperion has no other choice until now to stick to it whatever software or hardware patch it will take.
Notes about my presentation at the Alchimie 4 show : Comment 292 of 427ANN.lu
Posted by JoannaK on 30-Sep-2004 16:47 GMT
In reply to Comment 291 (Christophe Decanini):
Personally I've seen so many 'final fixes' during last 2-3 years that I tend to be a bit sceptic for this announcement. But *IF* it's real this time it's a good sing. If they can finally make hardware to be usable with Linux it opens up considerable new markets for Teron and AmigaOne sales.

But.. IMHO it'll needs a lot more proof than just one post on Ann.lu to make it happen. A confirmation from Eyetech would be a good start. So far, I have not seen any messages (on any site I do follow) that supports this claim.
Notes about my presentation at the Alchimie 4 show : Comment 293 of 427ANN.lu
Posted by Johan Rönnblom on 30-Sep-2004 16:51 GMT
In reply to Comment 288 (Agima):
Yeah! We should all go the Microsoft way, and backstab developers by
making their apps fail on new OS versions!

Really.. sometimes apps fail because they are making assumptions about
undocumented behaviour. Sometimes, although rarely, it's because
there's some real reason why some functionality is not provided (say,
copperlists). But usually when the documentation for an operating
system says that an API call will deliver certain results, that is a
"promise" to developers that they can rely on this. If they follow the
guidelines, their apps will work in the future.

Anyway, it's funny because in this case it doesn't (necessarily) seem
to be a design incompatibility in OS4. The discussion just popped up
because Thomas Frieden wrongly assumed that scsispeed does something
that it doesn't. Apparently the reason why Stéphane Guillard modified
it was some other problem, and not necessarily an "incompatibility by
choice".
Notes about my presentation at the Alchimie 4 show : Comment 294 of 427ANN.lu
Posted by Michal Schulz on 30-Sep-2004 17:54 GMT
In reply to Comment 278 (Georg Steger):
Actually the cacheline size is 32 bytes. So this hypothetical 512 byte sector Stephane mentioned may occupy 16 or 17 cache lines, depending on the alignment of the buffer.

However the issue remains. During DMA transfer it may happen that the 32-byte line is transferred again into cache just because other task has accessed it.
Notes about my presentation at the Alchimie 4 show : Comment 295 of 427ANN.lu
Posted by MikeB on 30-Sep-2004 17:58 GMT
In reply to Comment 287 (Christophe Decanini):
You told your side of the story while I believe you should have refrained, at least if you really didn't want to see me clarify things in here with regard to your "selected facts".

- One fact is that despite I have never been moderated here on ANN you want to see me banned.
- Another fact is that you have a personal interest, because of your involvement with a Pegaos dealer, for this platform to succeed.
- Another fact is that when I was asked to leave Genesi was involved here on ANN and later even hired its webmaster.
- Last but not least, you cannot stop insulting me and I believe that you would have moderated me if I would have shown the same behaviour against you.

> I did notice it but I guess talking about it will be considered by certain as
> "trolling".

I don't understand you guys, it has been known for a very long time that the AmigaOS4/Linux kernel/drivers would have to take into account the way the current ArticiaS chips operates. The AmigaOS4 team has clarified on multiple occasions that this would NOT be an issue with AmigaOS4.

How you guys could see positive news like this, that future AmigaOne boards won't have this problem with regard to standard unmodified Linux kernel/drivers as being so negative just beats me.
Notes about my presentation at the Alchimie 4 show : Comment 296 of 427ANN.lu
Posted by Stefan Burström on 30-Sep-2004 18:21 GMT
In reply to Comment 294 (Michal Schulz):
>However the issue remains. During DMA transfer it may happen that the 32-byte line is transferred again into cache just because other task has accessed it

Why would that happen? Remember that this has been the 'issue' since DMA was introduced with Z3 and so far this hasn't been a problem.

regards,
Stefan
Notes about my presentation at the Alchimie 4 show : Comment 297 of 427ANN.lu
Posted by Michal Schulz on 30-Sep-2004 18:27 GMT
In reply to Comment 219 (Stéphane Guillard):

For a given buffer to write to disk, with a given set of dirty cachelines in it (say the buffer is 512 bytes, say there are 3 32 byte cachelines that have to be flushed), according to what I explained, both methods will end up doing exactly the same cache flushes on the bus : 3 cacheline flushes happening in the middle of the DMA transfer with hardware coherency, and 3 cacheline flushes happening in the beginning of the DMA transfer with manual cache flushing, or software coherency. Guess which is faster ? Tell us a good tale here please.


Yes, mostly you're right. But mostly only. In your hypothetical case only 3 lines of cache are to be flushed. But if you do it with additional portion of code instead of use the snooping, the CPU has to invalidate all lines just because you (or Operating System if you prefer) doesn't know which portions of memory are cached and which parts of cache are dirty and needs to be flushed.

You may say that 16 lines of cache are not much, but in most cases there are much more than this 512 bytes readed. Especially because in case of small block transfers DMA becames totally useless as it performs as fast as PIO and the overhead known from PIO modes is than negligible.

Looking at the ATA driver from project I'm with:

Testing with a 512 byte, MEMF_FAST, LONG-aligned buffer.
Read from SCSI: 6182348 bytes/sec

Testing with a 4096 byte, MEMF_FAST, LONG-aligned buffer.
Read from SCSI: 36071833 bytes/sec

Testing with a 32768 byte, MEMF_FAST, LONG-aligned buffer.
Read from SCSI: 57291571 bytes/sec

Testing with a 262144 byte, MEMF_FAST, LONG-aligned buffer.
Read from SCSI: 57396428 bytes/sec

:)
Notes about my presentation at the Alchimie 4 show : Comment 298 of 427ANN.lu
Posted by Jacob on 30-Sep-2004 18:32 GMT
In reply to Comment 144 (Fabio Alemagna):
Fabio you look so retarded at times like this, give it a rest.
Notes about my presentation at the Alchimie 4 show : Comment 299 of 427ANN.lu
Posted by Michal Schulz on 30-Sep-2004 18:33 GMT
In reply to Comment 296 (Stefan Burström):
Imagine that the transfer buffer is allocated on 8 byte boundary and it unfortunatelly only 8 bytes of one of the 32 bit cache line belongs to the buffer. Rest is used elsewhere. Now imagine the line is flushed because the DMA transfer is to be occured. The tasks are switched during the DMA transfer (as it is to be - we have spare CPU time during the transfer and it may be used elsewhere) and the other task accesses the cache line which is shared between the sector buffer and some other thing. This very line of cache may then be fetched before it is overwritten with contents of the BM-transferred sector. As the HW cache coherency is not maintained, the CPU doesn't know that the cache line has been changed again. oops?
Notes about my presentation at the Alchimie 4 show : Comment 300 of 427ANN.lu
Posted by Fabio Alemagna on 30-Sep-2004 18:56 GMT
In reply to Comment 298 (Jacob):
> Fabio you look so retarded at times like this, give it a rest.

Oh, yessir, I will certainly follow your knowledged advice.
Anonymous, there are 427 items in your selection (but only 177 shown due to limitation) [1 - 50] [51 - 100] [101 - 150] [151 - 200] [201 - 250] [251 - 300] [301 - 350] [351 - 400] [401 - 427]
Back to Top