T.R | Title | User | Personal Name | Date | Lines |
---|
2254.1 | MIPS I Think... | FORTSC::CHABAN | Pray for Peter Pumpkinhead! | Tue Dec 01 1992 20:13 | 12 |
|
Is suspect that it is in response to some old news here in Silly-Con
valley. A local TV station here broke a story about how some
researcher at Stanford discovered a bug in the floating point hardware
on the high-end MIPS machines. For some reason, MIPS/SGI are
unaffected.
How much you wanna bet our friends at HP and Sun gave the TV station
the lead?
-Ed_in_Santa_Clara
|
2254.2 | This is a very public problem | POCUS::HUSTON | | Tue Dec 01 1992 22:31 | 8 |
| Unfortunately it's true but fortunately DEC is forthright about this
one. Two of my customers have already received notices that the
floating point processor on DECsystem 5000-240 and 5900's delivered
before August 1992 must be replaced. The memo from DEC is very clear -
there is no option but to replace, although you have little exposure if
you do no floating point operations.
dh
|
2254.3 | Conference Pointer | SDSVAX::SWEENEY | Patrick Sweeney in New York | Tue Dec 01 1992 22:58 | 1 |
| Disscused in SUBWAY::DIGITAL_INVESTING
|
2254.4 | | AOSG::NORDLINGER | To reach the unreachable STAR:: | Wed Dec 02 1992 07:39 | 3 |
| Wow, when it rains it poors, our luck has *got* to get better!
John
|
2254.5 | Rumors of reports | TPSYS::BUTCHART | TNSG/Software Performance | Wed Dec 02 1992 07:55 | 8 |
| According to a short article in the stock market section of the Boston
Globe, our stock dropped on a rumor of an analysts report that there
were problems in the Alpha chip. No such report ever materialized, and
the article quoted unnamed sources as saying the rumor originated from
some short sellers trying to trigger a drop so they could make some money
off Monday's run-up.
/Butch
|
2254.6 | | RUSURE::EDP | Always mount a scratch monkey. | Wed Dec 02 1992 08:31 | 6 |
| If there is a bug in some Mips chip, why the hell haven't engineers in
the Ultrix kernel group been informed so we can prepare a software
work-around?
-- edp
|
2254.7 | No wonder my benefits don't add up! | SPESHR::BENOIT | Life is just a cherra bowlies | Wed Dec 02 1992 08:48 | 7 |
|
We have had problems with floating point ever since the 11780 in
various machines. Engineering's response is that you can't test every
combination of every formula and every calculation. So I guess the
bottom line is - let the customer find the problem.
Peter
|
2254.8 | | SUBWAY::CATANIA | Mike C. �-� | Wed Dec 02 1992 10:00 | 6 |
| .3
How about a note number!!
- Mike
|
2254.9 | 10.320 | CSOADM::ROTH | Call off your goons, I give up! | Wed Dec 02 1992 10:27 | 5 |
| re: .3, .8
Try starting at note number above in DIGITAL_INVESTING.
Lee
|
2254.10 | | SPESHR::KEARNS | | Wed Dec 02 1992 13:08 | 8 |
|
re: .7
Can you back this up. Have we had problems in every single FPU? And
even if we did what would be your recommended solution given time to
market constraints?
- Jim K
|
2254.11 | | QUARK::LIONEL | Free advice is worth every cent | Wed Dec 02 1992 16:11 | 10 |
| The DECstation problem is not new; indeed it was known back in March. The
problem occurs with only certain boards; it is timing and data dependent.
I don't believe there's anything that ULTRIX could do to work around it.
A diagnostic is in place in manufacturing to identify bad boards (though
it is not 100% effective according to the DEC Fortran developer who created
it.) Customers have been made aware of the problem.
Anyway, this had nothing to do with the stock drop.
Steve
|
2254.12 | Decstation details | POCUS::HUSTON | | Wed Dec 02 1992 16:22 | 7 |
| .11 is right.
DS5000-240/5900 problem is not related at all to stock drop. It's not
entire line, just 40MHz cpu's. Sales Update Flash #735 points you to
note 4335.0 in DWT notes file for technical description.
dh
|
2254.14 | who's that looking over my shoulder? | ODIXIE::RHARRIS | The deerhuntermeistersupreme | Thu Dec 03 1992 08:01 | 11 |
| attention .13, if you read the FIRST sentence in the base note, it
states the word rumour.
Also, there is a sentence that states, please respond with FACTS.
me paranoid, nahh.
I am asking for facts, that's all.
bob
|
2254.15 | From yesterday's WSJ | VINO::FLEMMING | Have XDELTA, will travel | Thu Dec 03 1992 09:39 | 12 |
| The sale of some large blocks of Digital Equipment's stock also hammered its
share price, which ended down 1 1/4 to 33. Trading in Digital's shares was
halted for an order imbalance in the morning. Traders attributed the blocks
to the sale of shares by employees who bought the shares as part of a
periodic employee stock purchase program. The company said, however, that
the employee selling didn't account for the heavy volume of trading in its
stock. Digital emphatically denied unsubstantiated rumors of problems with
its Alpha computers, attributed to unnamed traders at the start of the session
[probably traders that had shorted the stock as a result of the runup Monday].
The company insisted there was no news to account for the volume and
attributed the move to two blocks offered at the opening. It also noted that
it was up 2 on Monday, also in heavy trading.
|