Author: Christophe Theron
Date: 11:01:31 11/07/01
Go up one level in this thread
On November 07, 2001 at 09:31:57, Steven Schwartz wrote:
>On November 06, 2001 at 23:27:06, Christophe Theron wrote:
>
>>On November 06, 2001 at 14:26:18, Steven Schwartz wrote:
>>
>>>On November 06, 2001 at 14:13:25, Christophe Theron wrote:
>>>
>>>>On November 06, 2001 at 13:59:41, O. Veli wrote:
>>>>
>>>>>Hello to all,
>>>>>
>>>>> I have seen the following info about Star Sapphire:
>>>>>
>>>>>Advanced SH7034 RISC processor (1M ROM and 256K RAM) and High Speed 20MHz energy
>>>>>saving CMOS TECHNOLOGY including Hash Tables.
>>>>>
>>>>> How comparable is the RISC processor with DragonBall used in Palms? Since we
>>>>>have already 33MHz Palms, and strong programs such as ChessTiger and ChessGenius
>>>>>for Palm, does Star Sapphire have a chance to claim to be world's strongest
>>>>>portable unit? Can RISC processor and/or 1M ROM add to its strength for the
>>>>>deficiency in speed? Thanks.
>>>>
>>>>
>>>>
>>>>I do not know how the SH7034 compares to the DragonBall.
>>>>
>>>>One thing I know is that you can FORGET ABOUT COMPARING THE MHz. It's
>>>>meaningless.
>>>>
>>>>The amount of ROM has no meaning either. I could use several Mb for Chess Tiger
>>>>for Palm if I wanted to, either for the program (code), the opening book, or the
>>>>hash tables. On some Palms I could use 16Mb.
>>>>
>>>>I do not know how the Star Sapphire will compare to Chess Tiger for Palm, but
>>>>I'm already almost certain that it will be washed away by PocketPC chess
>>>>programs like ChessGenius running on the iPaq.
>>>>
>>>>So its claim to be the world's strongest portable unit is bullshit.
>>>>
>>>>I have also received an ad today from ICD claiming its elo is between 2400-2550
>>>>elo. That must be expressed in BS elo (Baby School).
>>>> Christophe
>>>
>>>
>>>You may very well be right about the strength issue vs. the various
>>>palm-top computers. We won't have any firm answers for several months.
>>>It actually was not our intent to compare the Star Sapphire with those
>>>devices but rather chess computers. As I mentioned earlier, maybe using
>>>the word "dedicated" would have been a better choice.
>>>
>>>As far as the estimated rating... we based the 2400 to 2550 on Novag's
>>>estimate of 2550 and our belief that the computer was likely to be 50
>>>or more stronger than the 2350 we put on the Sapphire II (the Star
>>>Sapphire's predecessor).
>>>Steve (ICD/Your Move Chess & Games)
>>
>>
>>
>>The Sapphire II is rated 2012 by the SSDF. Maybe 2100 because the SSDF
>>subtracted 100 points to all the computers in the list, and while it was
>>justified for the top, it was less justified for the bottom of the list.
>>
>>IIRC the Sapphire II already had the same processor as the Star Sapphire.
>>
>>I have a hard time believing that the program would have improved so much that
>>it would now be 50, 100 or 150 elo points stronger.
>>
>>So I think your figures are overevaluated, even in USCF elo.
>> Christophe
>
>
>The ratings of chess computers have been a disaster for years. We
>have always tried to impart a little sanity into the situation by
>never listening to manufacturers. However, the Star Sapphire is
>about 6 to 8 weeks away, and we have not had a machine to test. So,
>we do the best we can.
>
>I recall that Enrique Irazoqui and Larry Kaufman used to tell us to
>add 180 points to the SSDF list to get USCF equivalents. If the
>older Sapphire II is rated 2012 by SSDF and if they subtracted 100
>points to make the ratings more in line with reality, and if we add
>180 to get the USCF version, then that would put the Sapphire II at
>2292.
>
>Also, the Sapphire II runs on a "RISC-style" H-8 Hitachi chip whereas
>the new Star Sapphire runs in a 68EZ328 true RISC. I am not a wizard
>on microprocessors, but I believe that the "EZ" chip is much faster.
>Maybe, some of the members of the group can enlighten me on the
>differences.
I don't understand how they can tell you such bullshit.
The 68EZ328 microprocessor from Motorola is also called the "DragonBall" EZ.
It's the microprocessor you can find in many older Palms including the PalmVx.
It is not the latest DragonBall. The latest is called DragonBall VZ, and it's in
the newest Palms, including the m505 and the latest Visors from Handspring.
The DragonBall VZ is faster than the EZ. The EZ runs at 20MHz by default, the VZ
defaults at 33MHz. Both are overclockable but I doubt that the Star Sapphire
will have an overclocked one.
The DragonBalls (68xx328) processors are not RISC processors at all. They have a
68000 core (you can find a 68000 in many older chess computers). I think it's
the first time in the world a company dares to say that the 68000 is a RISC. The
H8 was more RISC than the 68EZ328!
So basically they sell an unit which has an outdated processor, which happens to
be the processor used in 2 years ago Palms. That's not surprising, given that
they announce the Star sapphire since 2 years.
The current Palms have a faster processor, and this processor can be overclocked
(many users of Chess Tiger for Palm can make their Palm 2 to 2.5 times faster
with an overclocking utility).
Owners of a Palm will have a 2x to 4x faster computer than the Star Sapphire.
With this information in mind, I would not be surprised if the Star Sapphire was
actually weaker than the Sapphire II.
>Additionally, the older II had 160K ROM and 129K RAM whereas the
>Star has 1M ROM and 256K RAM.
>We used to quote 2300 to 2350 for the Sapphire II while Novag was
>quoting mid 2400s, and we sold quite a few, and I recall no complaints
>about the playing strength at all. Based upon program upgrades (which
>we can only assume are in place because the Star will be almost three
>years newer than the II), micro-processor upgrades, RAM and ROM upgrades,
>suggesting that the Star will be a minimum of 50 to 100 points stronger,
>in my opinion, is not a stretch at all. But we shall see.
It's not really a microprocessor upgrade. It looks rather like a downgrade.
More RAM and more ROM does not guarantee more strength.
50 to 100 elo points stronger would be a HUGE jump, especially on a slow
processor. And especially since we have not heard about the author of the
program since quite a long time.
I think you should be more careful about the manufacturer's claims on this
device.
Christophe
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.