|
|
|
|
|
|
| Author |
Message |
Atro Tossavainen
*nix forums Guru Wannabe
Joined: 22 Feb 2005
Posts: 131
|
 Posted: Sat Apr 23, 2005 7:52 pm Post subject:
Re: Dead Fuel
|
|
|
gheston@hiwaay.net (Gary Heston) writes:
| Quote: | Have you checked to verify that the fans are working? My general
experience with CPUs is that they either work or not, so if you're
seeing the boot diags work, the CPU may be ok, just overheating due
to a failed fan or missing duct.
|
Duh. No. The machine is making the normal noises and no unexpected
new ones, which I suppose is why I never even thought of the possibility.
Good point. Will be checking. Thank you.
--
Atro Tossavainen (Mr.) / The Institute of Biotechnology at
Systems Analyst, Techno-Amish & / the University of Helsinki, Finland,
+358-9-19158939 UNIX Dinosaur / employs me, but my opinions are my own.
< URL : http : / / www . helsinki . fi / %7E atossava / > NO FILE ATTACHMENTS |
|
| Back to top |
|
 |
chrisv
*nix forums Guru Wannabe
Joined: 24 Mar 2005
Posts: 124
|
| Posted: Tue May 03, 2005 12:54 pm Post subject:
Re: Xenon - Xbox360 getting embedded DRAM confirmed. NEC will manufacture the eDRAM graphics chip
|
|
|
Del Cecchi wrote:
| Quote: | Dell has of late been advertising a PC for USD
299, for example on the back cover of the supplement magazines that come in
the newspaper. I recently bought a 512MB, 3GHz/HT, 80GB SATA machine with a
19 inch LCD flat panel from Dell for USD 580 plus tax with free shipping.
|
And once you wipe it clean of all the pre-installed crap, it may
actually be a decent machine.  |
|
| Back to top |
|
|
Grant Schoep
*nix forums beginner
Joined: 21 Apr 2005
Posts: 11
|
| Posted: Wed May 04, 2005 1:17 am Post subject:
Re: Xenon - Xbox360 getting embedded DRAM confirmed. NEC will manufacture the eDRAM graphics chip
|
|
|
chrisv <chrisv@nospam.invalid> wrote in
news:f34f71dceh0864dt3ss6bhovntei1hsjj7@4ax.com:
| Quote: | Del Cecchi wrote:
Dell has of late been advertising a PC for USD
299, for example on the back cover of the supplement magazines that
come in the newspaper. I recently bought a 512MB, 3GHz/HT, 80GB SATA
machine with a 19 inch LCD flat panel from Dell for USD 580 plus tax
with free shipping.
And once you wipe it clean of all the pre-installed crap, it may
actually be a decent machine. 8)
|
Two things here first this part of the thread. Building boxes this cheap
isn't that hard, especialy for someone who can buy in bulk like Dell.
Looking at these machines they are using lots of integration graphics and
the like. Stuff just ends up being pretty cheap. They throw in a flat panel
to make it look a bit better. Dell has a good corner in on flat panel sales
right now so they probably get great deals. These boxes aren't great, but
there nice for the the parent that want's to get their kid off their
computer.
Oh, and one thing Dell is actually fairly good about, they don't preinstall
much crap. Sure, they got the MSN/Dellnet crap on the boxes. There's a few
other stupid things like screen savers and that but not really stuff that
affects much. hey, It is always easier to just reinstall windows anyways!
when you first get the box. |
|
| Back to top |
|
|
Luis Catulo
*nix forums beginner
Joined: 04 May 2005
Posts: 4
|
| Posted: Wed May 04, 2005 8:46 am Post subject:
Re: How to access L1 port (and logs) on FUEL ?
|
|
|
Jeff Long <long+news9@kestrel.cc.ku.edu> wrote in message news:<ob42k2-gr1.ln1@shibby.kicks-ass.net>...
| Quote: | Luis Catulo wrote:
Jeff Long <long+news9@kestrel.cc.ku.edu> wrote in message news:<f8fnj2-ahl.ln1@shibby.kicks-ass.net>...
Luis Catulo wrote:
Hi,
I have some problems with a SGI Fuel that I suspect come from a
damaged motherboard... The 5V aux keeps going over 6V and the machine
powers off!
Already tested with a new powersupply but the problem is still there!
On a newsgroup I read something about accessing the L1 logs... How can
I do this?
On ttyd1 I can access the normal prom but <ctrl-t> doesn't work as on
a Onyx2 machine (or similar) to access MSC...
I guess there is some kind of cable to connect to the L1?
I don't have much experience with Fuel machines...
Any help, cable pinout or anything, greatly appreciated!
If the system is running just use:
/usr/sbin/l1cmd log
Jeff Long
Thanks for your help Jeff!
But the machine isn't running... It starts making the diagnostics but
95% of the times it shuts down before I can log in... The other 5%
happens it gives me a couple of seconds to log...
Darn, I was hoping your machine ran a little while. You've about
exhausted my knowledge on this. I seem to remember reading somewhere or
seeing an SGI engineer using a linux laptop and a usb cable to do
low-level diagnostics. Perhaps installing one of the various versions
of the System Controller software on a linux laptop would allow this.
But that's about my only idea.
Ah yes, page 130 of the manual mentions a USB-B port used for system
maintenance.
Jeff Long
|
Hi,
Just "found" an 9 pin RS232 port right on the motherboard near the CPU module!
Any idea for its use (access L1???) ?
I connected there my terminal, but had no luck...
rgds,
Luis Catulo |
|
| Back to top |
|
|
Luis Catulo
*nix forums beginner
Joined: 04 May 2005
Posts: 4
|
| Posted: Wed May 04, 2005 9:09 am Post subject:
Re: o200 PROBLEM - PLEASE HELP
|
|
|
Hi,
You might have your O200 configured as a SLAVE unit and not a
MASTER...
If that is the case it would only start to boot if it was connected to
another O200 using a CRAYLINK cable. It is the MASTER machine that
sends the UP signal to the SLAVE.
There is a small DIP SWITCH that sometimes, specially after moving the
machine, "jumps" to the wrong position. This DIP controls whether the
machine is MASTER or SLAVE.
Check the system controller of the O200. You'll find there a group of
10 DIP's.
-If all 10 of them are on the same position, you found your problem!
That is the configuration for SLAVE.
DIP number 10 should be ON (or is it OFF?) to get a MASTER O200
machine...
Just look at DIP 10... If it is like the rest, reverse its position.
Of course, when you remove the system controller or if you access it
on the side panel, don't forget to remove the power cord from the
machine.
Regarding the starting LED sequence, it will be the same whether the
machine is salve or master, so you really have to look at the DIP's...
I don't know how much you know about O200, but the system controller
is a small board you can remove very easy... On front of the machine,
remove the plastic cover over the buttons/leds. You should see 2
screws. Remove them and pull the board out. The DIP's are there.
rgds,
Luis Catulo
stefan schuerch <stefan.schuerch@gmx.ch> wrote in message news:<4274011c$0$296$4d4ef98e@read.news.ch.uu.net>...
| Quote: | Hi Freddy,
I have exact the same problem with my o200. i' ve tested the machine
with minicom too. but no output. have you solve the problem?
stefan
Thanks very much for the advice Steve. I will try and get a cable now.
Freddy
Steve S. wrote:
bonehead wrote:
Steve,
No I do not have one, guess I could fabricate one though.
What functionality does the aux port provide?
Thanks
Freddy
Alot of times you can see console messages and any error messages
that
may be occuring. If you get nothing there I would say you have a bad
logic carrier / motherboard.
HTH
Steve
|
|
|
| Back to top |
|
|
Walter Mitty
*nix forums beginner
Joined: 04 May 2005
Posts: 16
|
| Posted: Wed May 04, 2005 3:22 pm Post subject:
Re: Xenon - Xbox360 getting embedded DRAM confirmed. NEC will manufacture the eDRAM graphics chip
|
|
|
Rob Stow wrote:
And after all that you still have to wonder if they need to make a
| Quote: | profit on the XBoxes. Selling at cost or even at a loss may be
acceptable to them if the games are where the expect to make their
profits. Bought an inkjet printer recently ?
|
This is not relevant : the fact is that they did sell the XBOX at a loss
in order to secure market etc. Yes, the inkjet analogy is a good one -
even if I didn't actually doubt it. |
|
| Back to top |
|
|
Guest
|
| Posted: Wed May 04, 2005 7:24 pm Post subject:
Re: V12 and antialiasing
|
|
|
zolo wrote:
| Quote: | I was reading up on SGI V12 graphics, which has a hardware
accumulation buffer. According to the SGI OpenGL
literature, the accumulation buffer enables such niceties as
full-scene antialiasing. However, it seems that none of the
Octane or Octane2 demos utilize antialiasing to any
noticeable degree. Is there a way to enable FSA for a given
app, or does it need to be specifically coded to utilize
that feature? Are there any free demos or apps that
demonstrate the antialiasing capabilities of the V12 hardware?
|
I believe most SGI workstations have hardware accumulation buffers,
even as far back as some iris 4D machines. You use the accumulation
buffer for antialiasing by rendering multiple passes of your scene with
slightly jittered viewpoints, and them blending them together. It needs
to be coded into the application, but is really quite easy. The
downside is it's slow (which I'm guessing is why you don't see it in
demos). The accumulation part is fast, but generally you have enough
trouble rendering your scene once at speed, let alone multiple times.
Cool tricks: using random jitters, adding random noise into the
accumulation (looks better), and using display lists for speed.
-dj |
|
| Back to top |
|
|
Herzog RTS
*nix forums beginner
Joined: 05 May 2005
Posts: 1
|
| Posted: Thu May 05, 2005 3:31 am Post subject:
Re: Xenon - Xbox360 getting embedded DRAM confirmed. NEC will manufacture the eDRAM graphics chip
|
|
|
"Doug" <pigdos@nospamcharter.net> wrote in message news:<3SSbe.22822$Jg7.20229@fe03.lga>...
| Quote: | I still don't understand how the hell MS is going to implement all this in
the Xbox and still charge under $300.00. Unless MS is willing take a huge
loss on every sale of the Xbox it doesn't seem possible to me that the Xbox2
is going to sell for less than $500.00.
|
IMHO,
It is not hard to understand--what MS is putting into Xbox360 is
really not that outrageous for a next-gen console that will cost $250
to $300 (that's the reported price-point for the basic model). It's
getting a relatively simple multi-core PowerPC CPU (only 2 or 3
cores). an ATI R500 graphics chip, an EDRAM module, standard GDDR3
memory, NO harddrive in basic model (saves alot of $$$$)
what *SONY* is having implemented in the *PS3* , now *that* is pretty
damn outrageous. the completely new, unproven CELL processor with *9*
cores (1 PowerPC + 8 auxillary processors). the brand new Rambus
Yellowstone / XDR memory. new expensive Blu-Ray disc drive. totally
new Nvidia GPU based on an architecture we've never seen (well to be
fair, Xbox360's ATI R500 is a similarly new architecture though)
what MS is doing with Xbox360 is pretty standard fair, and basicly
in-line with then-new consoles of the past, like Gamecube, Dreamcast,
Nintendo64, Playstation1, etc. All of which had technology that
seemed too expensive to be put into a console before they came out. |
|
| Back to top |
|
|
Andreas Eder
*nix forums beginner
Joined: 05 May 2005
Posts: 6
|
| Posted: Thu May 05, 2005 10:48 am Post subject:
Re: Xenon - Xbox360 getting embedded DRAM confirmed. NEC will manufacture the eDRAM graphics chip
|
|
|
Grant Schoep <matobinder@idcomm.abcremooveabc.com> writes:
| Quote: | chrisv <chrisv@nospam.invalid> wrote in
news:f34f71dceh0864dt3ss6bhovntei1hsjj7@4ax.com:
Del Cecchi wrote:
Dell has of late been advertising a PC for USD
299, for example on the back cover of the supplement magazines that
come in the newspaper. I recently bought a 512MB, 3GHz/HT, 80GB SATA
machine with a 19 inch LCD flat panel from Dell for USD 580 plus tax
with free shipping.
And once you wipe it clean of all the pre-installed crap, it may
actually be a decent machine. 8)
Oh, and one thing Dell is actually fairly good about, they don't preinstall
much crap. Sure, they got the MSN/Dellnet crap on the boxes. There's a few
other stupid things like screen savers and that but not really stuff that
affects much. hey, It is always easier to just reinstall windows anyways!
when you first get the box.
|
It would be better atill to install a real operating system instead of
reinstalling windows! :-)
'Andreas
--
Wherever I lay my .emacs, there's my $HOME. |
|
| Back to top |
|
|
keith
*nix forums beginner
Joined: 05 Jul 2005
Posts: 18
|
| Posted: Fri May 06, 2005 12:41 am Post subject:
Re: Xenon - Xbox360 getting embedded DRAM confirmed. NEC will manufacture the eDRAM graphics chip
|
|
|
On Wed, 04 May 2005 22:31:24 -0700, Herzog RTS wrote:
| Quote: | "Doug" <pigdos@nospamcharter.net> wrote in message news:<3SSbe.22822$Jg7.20229@fe03.lga>...
I still don't understand how the hell MS is going to implement all this in
the Xbox and still charge under $300.00. Unless MS is willing take a huge
loss on every sale of the Xbox it doesn't seem possible to me that the Xbox2
is going to sell for less than $500.00.
IMHO,
It is not hard to understand--what MS is putting into Xbox360 is
really not that outrageous for a next-gen console that will cost $250
to $300 (that's the reported price-point for the basic model). It's
getting a relatively simple multi-core PowerPC CPU (only 2 or 3
cores).
|
Oh, my. Two or three cores is "relatively simple"? I think you should
tell AMD and Intel this. IITC, the price=sheet for AMD's dual core chip
showed something north of $1500.
an ATI R500 graphics chip,
No difference.
| Quote: | an EDRAM module, standard GDDR3 memory,
|
<yawn> DRAM is DRAM.
NO harddrive in basic model (saves alot of $$$$)
$40, tops. Good grief!
<snip>
--
Keith |
|
| Back to top |
|
|
Paul Keinanen
*nix forums beginner
Joined: 27 May 2005
Posts: 21
|
| Posted: Fri May 06, 2005 5:31 am Post subject:
Re: Xenon - Xbox360 getting embedded DRAM confirmed. NEC will manufacture the eDRAM graphics chip
|
|
|
On Thu, 05 May 2005 22:41:39 -0400, keith <krw@att.bizzzz> wrote:
| Quote: | Oh, my. Two or three cores is "relatively simple"? I think you should
tell AMD and Intel this. IITC, the price=sheet for AMD's dual core chip
showed something north of $1500.
|
That is the list price, which can be quite arbitrary, but the price
paid by a very large customer reflects much more accurately the
production costs. The number of defect chips on a wafer is directly
proportional to the chip area (complexity).
Assuming that the dual core chip would be twice as large as a single
core chip, the number of dual core chips obtained from a wafer would
be one half of single core chips, thus the price would have to be
twice to cover the wafer costs. Further assuming that the yield for
the single core chips would be as low as 50 %, then the twice as big
dual core yield would only be 25 % and the dual core chip price would
again have to be doubled or be 4 times the single chip price to cover
the wafer costs.
However, the dual core chips have some circuitry in common, so the
area would not be twice as large for a double core chip and assuming a
higher yield for a single core chips (say 90 %), the yield for a twice
as large chips would still be quite acceptable (81 % in this case).
Thus, the manufacturing costs would only be about twice as high as a
single core chip.
From yield point of view, it makes much more sense to make multiple
simple cores on chips than a extremely complex single core with the
same chip area as the multicore chip. If there is a single defect in
the complex chip, the whole chip is lost. However, if a single core in
a multicore chip is defect, the defect chip is disabled and sold as a
single core chip. I would be very surprised, if any semiconductor
maker would design 3 core chips, but most likely these are originally
4 core chips, with one failed core, which are sold as 3 core chips
usually to some big customers at a greatly reduced price.
In the days of the first 16 kbit DRAMs, Intel apparently had problems
with making perfectly good 16 kbit chips, since they themselves made
large 1 Mbyte core memory rack replacement modules (with ECC) using
more than a thousand "8 kbit" chips, which in fact were failed 16 kbit
chips.
Also remember the 386SX/387 issue, in which at least with some early
mask versions of the 386SX it was in fact a 386DX mask with the
floating point processor disabled, apparently due to defects in the
FPP area, although later 386SX versions were made without the FPP in
the mask for the low cost market.
I have no idea, if AMD is using a similar strategy for these multicore
chips, but at least it would be logical to do so.
Paul |
|
| Back to top |
|
|
keith
*nix forums beginner
Joined: 05 Jul 2005
Posts: 18
|
| Posted: Fri May 06, 2005 10:49 am Post subject:
Re: Xenon - Xbox360 getting embedded DRAM confirmed. NEC will manufacture the eDRAM graphics chip
|
|
|
In article <si0m71dbr7a7l4i0tdliuc2b63ddbl4rqr@4ax.com>,
keinanen@sci.fi says...
| Quote: | On Thu, 05 May 2005 22:41:39 -0400, keith <krw@att.bizzzz> wrote:
Oh, my. Two or three cores is "relatively simple"? I think you should
tell AMD and Intel this. IITC, the price=sheet for AMD's dual core chip
showed something north of $1500.
That is the list price, which can be quite arbitrary, but the price
paid by a very large customer reflects much more accurately the
production costs.
|
Baloney. No one sells silicon, particularly something with as much IP
in it as a processor, for anything close to "production costs", to
*ANYONE*. Selling at close to cost isn't a good way to pay the note on
a multi-billion dollar fab, a hundred million in engineers salaries,
while leaving enough to keep the stock holders happy.
| Quote: | The number of defect chips on a wafer is directly
proportional to the chip area (complexity).
|
Maybe, sorta. Good enough for a first swag, anyway. Large caches give
opportunities to change these simple rules somewhat. But that has
nothing to do with the issue at hand.
<snip>
| Quote: | From yield point of view, it makes much more sense to make multiple
simple cores on chips than a extremely complex single core with the
same chip area as the multicore chip. If there is a single defect in
the complex chip, the whole chip is lost. However, if a single core in
a multicore chip is defect, the defect chip is disabled and sold as a
single core chip. I would be very surprised, if any semiconductor
maker would design 3 core chips, but most likely these are originally
4 core chips, with one failed core, which are sold as 3 core chips
usually to some big customers at a greatly reduced price.
|
Oh, so now they're going to increase the chip size yet again to add a
fourth core (and all the interconnect) knowing that they're going to
throw it away? Yikes! The cores we're talking about aren't a sq. mm.
| Quote: | In the days of the first 16 kbit DRAMs, Intel apparently had problems
with making perfectly good 16 kbit chips, since they themselves made
large 1 Mbyte core memory rack replacement modules (with ECC) using
more than a thousand "8 kbit" chips, which in fact were failed 16 kbit
chips.
|
Sure, partial-good strategies have been used for memories for a *long*
time, but memory <> processors. BTW, memory makers now use row/column
redundancy with fuses blown during test to replace defective
rows/columns.
| Quote: | Also remember the 386SX/387 issue, in which at least with some early
mask versions of the 386SX it was in fact a 386DX mask with the
floating point processor disabled, apparently due to defects in the
FPP area, although later 386SX versions were made without the FPP in
the mask for the low cost market.
I have no idea, if AMD is using a similar strategy for these multicore
chips, but at least it would be logical to do so.
|
Could be, but it's irrelevant to the issue at hand.
--
Keith |
|
| Back to top |
|
|
Evgenij Barsukov
*nix forums beginner
Joined: 06 May 2005
Posts: 1
|
| Posted: Fri May 06, 2005 12:58 pm Post subject:
Re: Xenon - Xbox360 getting embedded DRAM confirmed. NEC will manufacture the eDRAM graphics chip
|
|
|
Paul Keinanen wrote:
| Quote: | On Thu, 05 May 2005 22:41:39 -0400, keith <krw@att.bizzzz> wrote:
Oh, my. Two or three cores is "relatively simple"? I think you should
tell AMD and Intel this. IITC, the price=sheet for AMD's dual core chip
showed something north of $1500.
That is the list price, which can be quite arbitrary, but the price
paid by a very large customer reflects much more accurately the
production costs. The number of defect chips on a wafer is directly
proportional to the chip area (complexity).
|
This is correct. Howver, the number of defect chips is about 5% after
production is decently ramped up.
| Quote: | Assuming that the dual core chip would be twice as large as a single
core chip, the number of dual core chips obtained from a wafer would
be one half of single core chips, thus the price would have to be
twice to cover the wafer costs.
|
Here is where the logic got screwed. If you have defects in 5% of chips
before doubling the complexity, and even if we follow strictly your
assertion that defects double, it will mean we have now 10% defects, and
by no means have half of the good chips per wafer - that means instead
of 95% good chips we now have 90% good chips.
This have increased the price by whole 5%.
Regards,
Evgenij |
|
| Back to top |
|
|
keith
*nix forums beginner
Joined: 05 Jul 2005
Posts: 18
|
| Posted: Fri May 06, 2005 3:58 pm Post subject:
Re: Xenon - Xbox360 getting embedded DRAM confirmed. NEC will manufacture the eDRAM graphics chip
|
|
|
In article <d5gace$rta$1@grapevine.wam.umd.edu>, foo@bar.invalid
says...
| Quote: | In comp.sys.ibm.pc.hardware.chips Keith R. Williams <krw@att.bizzzz> wrote:
In article <si0m71dbr7a7l4i0tdliuc2b63ddbl4rqr@4ax.com>,
keinanen@sci.fi says...
Also remember the 386SX/387 issue, in which at least with some early
mask versions of the 386SX it was in fact a 386DX mask with the
floating point processor disabled, apparently due to defects in the
FPP area, although later 386SX versions were made without the FPP in
the mask for the low cost market.
Could be, but it's irrelevant to the issue at hand.
Can't be. It may still be irrelevant to the issue at hand, but the
facts here are incorrect.
|
Ok. ;-)
| Quote: | The difference between 386SX and 386DX is the pinout, not the FPU.
386's never had integrated FPU's.
The difference between 486SX and 486DX is the presence/absence of an
FPU.
For bonus points, figure out what a 487SX was without looking it up in
reference texts.
|
A 486DX with the integer CPU disabled?
--
Keith |
|
| Back to top |
|
|
Paul Keinanen
*nix forums beginner
Joined: 27 May 2005
Posts: 21
|
| Posted: Fri May 06, 2005 4:18 pm Post subject:
Re: Xenon - Xbox360 getting embedded DRAM confirmed. NEC will manufacture the eDRAM graphics chip
|
|
|
On Fri, 6 May 2005 08:49:40 -0400, Keith R. Williams <krw@att.bizzzz>
wrote:
| Quote: | In article <si0m71dbr7a7l4i0tdliuc2b63ddbl4rqr@4ax.com>,
keinanen@sci.fi says...
On Thu, 05 May 2005 22:41:39 -0400, keith <krw@att.bizzzz> wrote:
Oh, my. Two or three cores is "relatively simple"? I think you should
tell AMD and Intel this. IITC, the price=sheet for AMD's dual core chip
showed something north of $1500.
That is the list price, which can be quite arbitrary, but the price
paid by a very large customer reflects much more accurately the
production costs.
Baloney. No one sells silicon, particularly something with as much IP
in it as a processor, for anything close to "production costs", to
*ANYONE*. Selling at close to cost isn't a good way to pay the note on
a multi-billion dollar fab, a hundred million in engineers salaries,
while leaving enough to keep the stock holders happy.
|
Of course you have to include the investment amortisation into the
"production cost" at some expected production volume. However, the
early "list price" charged from early adaptors can be quite arbitrary.
| Quote: | The number of defect chips on a wafer is directly
proportional to the chip area (complexity).
Maybe, sorta. Good enough for a first swag, anyway. Large caches give
opportunities to change these simple rules somewhat. But that has
nothing to do with the issue at hand.
|
This again depends if there are only core specific caches or just a
common cache on the chip.
| Quote: | From yield point of view, it makes much more sense to make multiple
simple cores on chips than a extremely complex single core with the
same chip area as the multicore chip. If there is a single defect in
the complex chip, the whole chip is lost. However, if a single core in
a multicore chip is defect, the defect chip is disabled and sold as a
single core chip. I would be very surprised, if any semiconductor
maker would design 3 core chips, but most likely these are originally
4 core chips, with one failed core, which are sold as 3 core chips
usually to some big customers at a greatly reduced price.
Oh, so now they're going to increase the chip size yet again to add a
fourth core (and all the interconnect) knowing that they're going to
throw it away? Yikes! The cores we're talking about aren't a sq. mm.
|
Those 4 core chips are made for customers that really need them and
are willing to pay the premium price for them.
These customers might not be willing to pay anything for chips with
one or two defective chips. Instead of scrapping these chips, there
might be others that are paying something for chips with 2 or 3
functional cores. It makes perfectly sense for the chip maker to sell
these chips instead of scrapping them.
However, if the demand for three core chips exceeds the amount of
partially faulty four core chips, only then it makes sense to produce
chips with a mask for only three cores. This threshold depends how the
yield for the 4 core chips is developing and how fast the demand for 3
core chips is growing.
| Quote: | In the days of the first 16 kbit DRAMs, Intel apparently had problems
with making perfectly good 16 kbit chips, since they themselves made
large 1 Mbyte core memory rack replacement modules (with ECC) using
more than a thousand "8 kbit" chips, which in fact were failed 16 kbit
chips.
Sure, partial-good strategies have been used for memories for a *long*
time, but memory <> processors.
|
With a sufficiently large number of cores on a chip, we are
approaching a similar tradeoff.
Paul |
|
| Back to top |
|
|
Google
|
|
| Back to top |
|
|
|
|
The time now is Tue Dec 02, 2008 5:42 am | All times are GMT
|
|
Loans | Mortgage Calculator | Air Force One Shoes | El libro de los nombre | Loans
|
|
Copyright © 2004-2005 DeniX Solutions SRL
|
|
|
|
Other DeniX Solutions sites:
Unix/Linux blog |
electronics forum |
medicine forum |
science forum |
|
|
Privacy Policy
|
Powered by phpBB © 2001, 2005 phpBB Group
|
|
FAQ
Search
Memberlist
Usergroups
Register
Profile
Log in to check your private messages
Log in
Forum index
»
*nix
»
SGI/IRIX
»
hardware
