SPIDER software update 3.16 and benchmarks

[DJS]

Elbox Computer is happy to announce availability of a new version of the spider.device driver for the Poseidon stack for the SPIDER and SPIDER II High-Speed USB 2.0 controllers. The new software is available free of charge for all registered users of the SPIDER and SPIDER II High-Speed USB 2.0 controllers.

The SPIDER software has been reworked in order to achieve the maximum possible performance, especially when SPIDER is used in the Mediator systems equipped with the Radeon 9200 graphic card. The large memory on the Radeon cards enables expansion of the DMA buffer used by PCI cards in the busmaster mode. The spider.device ver. 3.16 allows setting the DMA buffer size for the Spider EHCI controller from 40 kB (as used in the previous spider.device versions) up to 2 MB. The size of the Spider DMA buffer is defined by the Envarc:Mediator/SpiderBuf variable.

Go to the tests page to learn the benchmarks for the SPIDER card with the DMA buffer set to 2 MB. The tests were performed on Amiga 1200 and Mediator 1200 TX with the most popular turbo cards. A 4 GB USB memory stick was used in the tests.

The Hi-Speed USB mass storage devices connected to the SPIDER USB controller reach 3 to 5 MBytes/s (depending on the turbo card), which is fast enough to allow comfortable work with the USB devices like memory sticks, digital cameras and Flash card readers. In case of hard drives we still recommend connecting them to the FastATA controllers, which allow twice as fast transfers (5 to 10 MBytes/s) in the same turbo card configuration (see the FastATA MK-III benchmarks).

Now we are working on the OS4 native USB 2.0 drivers for the Sirion USB stack. They will be included in the next SPIDER update.

[Zac67]

Too bad Chris' Poseidon stack doesn't support Spider any more...  :rtfm:

[platon42]

Too bad Chris' Poseidon stack doesn't support Spider any more...

Right. I suppose Elbox have created these Benchmarks with their pirated Poseidon copy. And the Deneb is still nearly twice as fast :-P And a buffer of 2 MB Pseudo-DMA Buffer is nice for creating artificial latency :-) And still, AFAIK the Spider does not support USB 1.1 devices behind USB 2.0 hubs (no split transactions), nor does it support IO queuing needed for serial and ethernet drivers, nor ISO transfers nor Realtime ISO as required for USB Audio... I could go on and on... :-P

[adolescent]

And the Deneb is still nearly twice as fast  :-P

On the A1200? :crazy: Funny that you're so concerned about a product that doesn't have any support that you continuously troll any Elbox news.

[amigadave]

@adolescent,

Yes, it is too bad that there is still such bad blood between Chris and Elbox and everytime Chris attacks them, he hurts his own self, as no one wants to be included in the middle of their argument.  It would be better for Chris if he just forwarded his observations and test results to a friend and had them post the comments objectively and without emotion, for Chris.  That would serve his purpose without making him look so vindictive.

[platon42]

It is *my* problem that users still buy new Spiders and then end up sending me several mails asking for a keyfile. It is not *your* problem with people pirating *my* product. You are not affected by Elbox-Fanboy-Trolls ranting on the press release of the Deneb (e.g. on amiga-news.de).

You wouldn't believe how many users write for registration and tell me they didn't know that Poseidon was no longer available. So yes, I post on threads trying to promote the Spider, a product which does not have software stack available for over a year now, but that obviously doesn't keep Elbox from selling this dud, leaving the buyer in the rain (and forced to get a pirated keyfile) and me with flame mails. So please, don't try to lecture me on morals here.

Comparing the benchmarks with the Deneb indeed was uncalled for. I'm sorry for that.

[amigadave]

It is not about morals that I am concerned.  I was just stating my opinion that it would be in your better interest to have your attacks on Elbox come from another person, or ghost as it were.

Personal attacks are generally frowned upon and I know that you have been attacked in return by the fan-boys of Elbox in the past, but you have also been attacked, or at least disowned by other Amiga users who don't know, or understand the whole story, no matter how many times you tell it.

You are obviously free to take your own approach to this situation you find yourself in, as I am sure you don't want the advice of anyone else.

I have no such fan-boy alliance with anyone.

[platon42]

Well, I prefer not to hide behind anonymous nicks or fake accounts or post cracking instructions on mailing lists using anonymous proxies with fake email addresses. I do have my own attitude and I show it. I'm not as schizo as the Elbox press department. I don't have to please all people out there.

Aimee Mann once sang: "I've got love and anger, they come as a pair. You can take your chances, but buyer beware."

[mboehmer_e3b]

On the A1200?

Well, just a word on this issue.

Guess why the Mediator 1200 has been choosen as comparison platform? Because the DENEB can "only" work there behind a (Fast-)Zorro II bridge, with all penalties created by the necessary synchronisation job inside this bridge, while the Spider resides in this case directly at the A1200 local CPU bus.

Nevertheless, if you compare the values for the 1230 CPU board, you will see that DENEB still reaches 80% of the Spider data transfer values, and this in a direct comparison between a local 50MHz CPU bus connected device (Spider) and an "outdated" Fast-Zorro II device (DENEB) connected via a 7.14MHz synchronous bus.

Amazing, as I think.

For whatever reason I cannot find any benchmarks for A4000 Zorro III based Mediators in this list.
Apparently this part of the Mediator line has been "forgotten", as the Radeon gfx card (providing the fast fake DMA buffer) simply does not work there due to the missing 3.3V power supply, and never will work there.

If one is making comparisons, one should make them under same conditions on the same hardware configuration, and not compare apples with pears (or, in other words, a "transfer rate up to 264MB/s" bus like the Mediator is advertised against a 3.5MB/s legacy bus).

Just my opinion, your one may vary, but hey, everyone may express his own opinion here. We are finally not on the Mediator mailing list where censorship still is used to silence people with unpleasent point of views :-)

Michael

[redrumloa]

On the A1200?  Funny that you're so concerned about a product that doesn't have any support that you continuously troll any Elbox news.

I think you are looking at this the wrong way. Elbox is promoting hardware based on software they do not have a license for. Promoting piracy is a good thing? Just my personal opinion.

[Gav]

I think its very stupid of elbox to still sell this product when a new buyer is unable to use it now..
I would buy one but since i cant use it there is no point,i can see why Chris is angry at elbox.

[Akiko]

For whatever reason I cannot find any benchmarks for A4000 Zorro III based Mediators in this list.
Apparently this part of the Mediator line has been "forgotten", as the Radeon gfx card (providing the fast fake DMA buffer) simply does not work there due to the missing 3.3V power supply, and never will work there.

@mboehmer_e3b

Maybe you missed this paragraph from the original Elbox Radeon press announcment

Compatibility
The Radeon Mediator drivers are designed for all MEDIATOR models. However, as most of the Radeon 9200 PCI series cards require the 3.3V auxiliary PCI power supply, they will work out of the box only in the Mediator 1200 TX and Mediator 1200 SX busboards with the ATX power supply.

PCI Radeon cards may also be used in other Mediator models, but it will require installing the Radeon card in the special PCI adapter which provides an auxiliary 3.3V power supply. Elbox is planning to produce a PCI 3.3V power adapter riser. It will include the bracket for the low profile Mobility Radeon 9200 card model available in Elbox Online Store.

[adolescent]

@mboehmer_e3b

Chris brought up the Deneb benchmark, Elbox didn't.    

Of course the Deneb is a great product.  Hopefully there will be enough made to satisfy the demand (and hopefully the US dollar recovers a bit so it becomes affordable for us in the states.  :-()

@red

Where in this post did Elbox promote piracy?  For the people that already registered Poseidon this is a new driver (and an announcement that OS4 drivers are being developed.).

[Crumb]

@adolescent

My A4000 has a Mediator. It is not an exclusive A1200 product, don't you know?

[TNovosel]

I have Mediator inside my A4K, and I also have SpiderUSB with spider.device, but that card never been "active". After I send email to Elbox, asking for software for my Spider card, Elbox replay with link to Chris page for ordering Posedion stack.
And now, after 3 years I still don't have any use of that Spider card.

If any other amiga user wanna useless card inside Amiga, go and buy Spider from Elbox. :crazy: