hb136@leah.Albany.Edu (Herb Brown) (07/06/89)
The following was posted on GEnie (file #6588). Several members of our local
Amiga user group are contemplating purchasing a Xetec controller. We would
appreciate any feedback concerning this post from current Xetec controller
owners. Thanks in advance.....Herb Brown Math Dept, The Univ at Albany
hb136@leah.albany.edu
-----------------------------------------------------------------------------
Are Xetec Controllers that Fast?
Consumer Alert
Don't believe everything you read! Case in point... In the July 1989
(Volume 5, Number 7) issue of Amiga World Bob Ryan, the magazine's technical
editor, did a comparative review of SCSI Controllers and Hard Drives for the
Amiga A-2000 and A-500. As a result of the testing Mr. Ryan rated the
Xetec FastCard/FastTrack products as the fastest with quotes like `leaves
everything in the dust' and `the fastest hard drive I have ever seen on the
Amiga.' The reviewer's recommendation about the Xetec FastCard for the
A-2000: `it is the best buy in the group.' And in the Amiga A-500 article
Mr. Ryan's recommendation was: `It is an amazing piece of hardware and I
recommend it highly...' Both articles included test results from the
PerfTest and DPerf 2.0 tests confirming that the Xetec products were indeed
the fastest products tested, with read speeds in excess of 750K bytes/sec.
If Mr. Ryan did in fact test the Xetec devices it is incomprehensible
that he could have written such a glowing report. We submit the following
for your consideration:
1> Xetec's own literature states that their Maximum Transfer Rate is 597K
bytes/sec. and their average read speed is 322 Kbytes/sec.
2> Simple mathematics indicate that it is impossible for an 8 bit bus
controller, like the Xetec to exceed 600K bytes/sec.
3> Common sense dictates that any product that provides such astounding
results from a non-innovative design should be subjected to additional
4> DPaint III would not load (every attempt to load resulted in a GURU) with
the Xetec controller. The file was verified as good and loaded when
another manufacturer's controller was used.
5> Large DigiPaint and Diamond files were consistently corrupted when saved
using the Xetec controller. The files were verified as corrupt with
DiskX.
And now, the rest of the story....
While it is true that the Xetec controllers did provide test results
that were stated in the Amiga World articles, Xetec (either accidentally or
on purpose) designed a MAJOR BUG into their controller so that it would too
well. It works like this: all of the disk performance test programs have
several things in common; they all use standard AmigaDOS Save/Load routines,
they all store successively larger files and none of them test for data
integrity. The BUG in the Xetec products prevents them from doing sequential
reads or writes of more than 128K of data, and it does so without reporting
any error of any kind to indicate that the file was not written or read
properly. This allows the Xetec products to read (or write) only the first
128K of a 1Meg file and still seemingly complete the test, and very quickly
at that!
Long sequential reads and writes were not allowed prior to the advent
the FastFileSystem introduced with AmigaDOS 1.3. Since the introduction of
FFS, long reads and writes are still not commonly encountered, but are by no
means a rarity and a SCSI controller should be able to handle them. DPaint
III, for example, would not load with the Xetec controller because during
loading it requests a load of greater that 128K.
The real insidious part of the problem with the Xetec controllers is
not that they read or write corrupt files, it is that they can do so without
reporting any errors! It is quite possible that an unsuspecting user could
spend many months building a data base without knowing that it had been
corrupted, and because of the lack of warnings, all of his backups would be
corrupted as well.
We're not sure which is worse, the fact that a major manufacturer could
have released a product with so little product testing that such a major
defect could have gone un-noticed, or the fact that Mr. Ryan could have
`overlooked' these major product defects and highly recommended the Xetec
controllers in a publication with the reputation, respect and readership
that Amiga World has. The bottom line is:
a) Many unsuspecting Amiga owners will buy the Xetec controller and
rely on it because of the Amiga World reviews.
b) The owners will experience problems with their systems.
c) And, as usual, the entire Amiga community will take the blame
for the unreliability. It is our sincere hope that this report
and be distributed quickly enough to prevent this.
In case you were wondering, none of the C Ltd. SCSI controllers were
allowed to compete in the test because according to Amiga World, they were
only testing Autobooting controllers and at the time of the test our
Autoboot was not yet finished. It is now, however, and the chart below
shows how the C Ltd. controllers should have been rated.
(Note: The tests of the C Ltd. controllers were using a Quantum hard drive,
as used by Amiga World, and are an average of several tests. Note SOME Manufacturers
were unsuccessful in using this HARDdrive -- We supports MOST every drive made.)
DPerf 2.0 tests results
256K 64K 32K Auto
Manufacturer Product Read Write Read Write Read Write Boot DMA Retail
C Ltd. Kronos/2000 623K 417K 548K 347K 430K 281K YES NO $299.95
MicroBotics HardFrame 590K 352K 605K 330K 570K 290K YES YES $329.95
Commodore A-2090A 584K 337K 537K 270 463K 247K YES YES $399.95
Supra Corp. SupraDrive 391K 267K 381K 264K 315K 234K YES YES $299.95
C Ltd. SCSI/2000 297K 171K 291K 167K 265K 157K NO NO $139.95
I.V.S. Trump Card 292K 230K 293K 219K 291K 214K YES NO $189.95
GVP Impact A2000 275K 154K 270K 154K 265K 157K YES NO $360.00
Exp. Tech. Flash!Card 55K 50K 55K 50K 55K 50K YES YES $299.95
---------------------------------------------------------------------------alh@hprmokg.HP.COM (Al Harrington) (07/06/89)
There is a discussion going on a local BBS here about Amiga World. Everyone has come to the conclusion that it is by far the most useless Amiga magazine out there. Articles like this seem to make that statement true. +--------------------------------------------------------------------------+ | -Al Harrington /// "The next thing I say to you will | | /// be true. | | alh@hprmo.HP.COM \\\/// The last thing I said was false." | | ..{hplabs,hp-sde}!hprmo!alh \XX/ -Devo | +--------------------------------------------------------------------------+ | My comments in no way reflect the views or opinions of HP | +--------------------------------------------------------------------------+
janhen@kunivv1.sci.kun.nl (Jan Hendrikx) (07/06/89)
In article <1889@leah.Albany.Edu> hb136@leah.Albany.Edu (Herb Brown) writes: |The following was posted on GEnie (file #6588). | |Are Xetec Controllers that Fast? [...] | The BUG in the Xetec products prevents them from doing sequential |reads or writes of more than 128K of data, [...] | | Long sequential reads and writes were not allowed prior to the advent |the FastFileSystem introduced with AmigaDOS 1.3. Since the introduction of |FFS, long reads and writes are still not commonly encountered, but are by no |means a rarity [...] This shows us how the people at AmigaWorld, the people at Xetec, and the (unknown) poster of this warning all forgot about MaxTransfer in the mountlist. -Olaf
rachamp@mbunix.mitre.org (Richard A. Champeaux) (07/07/89)
In article <13240030@hprmokg.HP.COM> alh@hprmokg.HP.COM (Al Harrington) writes: >There is a discussion going on a local BBS here about Amiga World. >Everyone has come to the conclusion that it is by far the most useless >Amiga magazine out there. Articles like this seem to make that >statement true. > >| -Al Harrington /// "The next thing I say to you will | If I remember right, the AmigaWorld article didn't say that the Xetec controller was the best. They said that if you are willing to spend the money, then the HardFrame 2000 was the one to get because it had the best overall times. They said that the Xetec had the distinction of having the fastest transfer rate for 256k buffers. (probably due to the bug that was mentioned) The Xetec was their second choice. I wouldn't say that AmigaWorld is useless. Sure, it may not be high on content, and definitely not a hacker's magazine (Amiga Transactor is the the one to get if you're looking for a hacker's magazine.), but AmigaWorld does have a lot of advertisements, and that's exactly why I like it. It's kind of like Computer Shopper, which is the best magazine to get if you're looking for something like drives, monitors, and printers. About 350 pages of the 400 pages Computer Shopper has are advertisements. Although the ratio isn't as big in AmigaWorld, it does seem to attract more advertisers than the other amiga magazines. Rich Champeaux (rachamp@mbunix.mitre.org)
tony@hp-sdd.hp.com (Tony Parkhurst) (07/11/89)
In article <1889@leah.Albany.Edu> you write: >4> DPaint III would not load (every attempt to load resulted in a GURU) with > the Xetec controller. The file was verified as good and loaded when > another manufacturer's controller was used. I am unable to load DPaintIII with my StarDrive controller, but DPaintII loads just fine. What is this flakyness??? Do you suppose there is something funky about DPaintIII? Anybody have ideas about this? -- Tony -- Tony Parkhurst ( tony@hp-sdd.HP.COM ) "Is this Hell? Or is this Texas?" "Both" -- Heinlein, _J_O_B: _A _C_o_m_e_d_y _o_f _J_u_s_t_i_c_e