A computer components & hardware forum. HardwareBanter

If this is your first visit, be sure to check out the FAQ by clicking the link above. You may have to register before you can post: click the register link above to proceed. To start viewing messages, select the forum that you want to visit from the selection below.

Go Back   Home » HardwareBanter forum » General Hardware & Peripherals » General
Site Map Home Register Authors List Search Today's Posts Mark Forums Read Web Partners

L2 cache question



 
 
Thread Tools Display Modes
  #1  
Old February 13th 06, 03:42 AM posted to alt.comp.hardware
external usenet poster
 
Posts: n/a
Default L2 cache question

Hi,
I buried my cpu, just tinkering to raise temp to my custom cooling. My
512kb l2 cahe p4 died at same spot as an intel p2 i had..(close to it).
Multithreading better than ever? Anyhow, my friend had the 1mb l2 cache p4
on same 533 bus, it did alot of work more before same choking. Prioritizing
didn't seem to exist! (this is only after truly conquering heat to an
extreme) the only drawback is the cpu killed his mobo and was very hot
runner at loaded. Is the 800FSB better with 1mb l2 cache or does it matter?
I like the 533 and only want increase of L2, but I don't know all the facts,
and can't find an end users result.Any advice appreciated.


  #2  
Old February 13th 06, 04:28 AM posted to alt.comp.hardware
external usenet poster
 
Posts: n/a
Default L2 cache question

On Mon, 13 Feb 2006 02:42:57 GMT, "bgd"
wrote:

Hi,
I buried my cpu, just tinkering to raise temp to my custom cooling. My
512kb l2 cahe p4 died at same spot as an intel p2 i had..(close to it).
Multithreading better than ever? Anyhow, my friend had the 1mb l2 cache p4
on same 533 bus, it did alot of work more before same choking. Prioritizing
didn't seem to exist! (this is only after truly conquering heat to an
extreme) the only drawback is the cpu killed his mobo and was very hot
runner at loaded. Is the 800FSB better with 1mb l2 cache or does it matter?
I like the 533 and only want increase of L2, but I don't know all the facts,
and can't find an end users result.Any advice appreciated.


I assume the first couple sentences you wrote are important,
but they don't make a lot of sense within the context of the
next lines.

It would be good to get your cooling ironed out, what good
is custom cooling if you kill cpus? Plain old stock cooling
has to be better than that.

The benefit of a larger L2 cache depends on the task. Some
respond better than others and if you have a particular
need, certain tasks that stress the CPU most then those
would have to be considered. Generally speaking you would
want the 800 pseudo-MHz (since it's quad-pumped, 200MHz
clocked) for the best performance, or rather, having the
memory bus faster than 133MHz / DDR266 as well.
  #3  
Old February 13th 06, 05:38 AM posted to alt.comp.hardware
external usenet poster
 
Posts: n/a
Default L2 cache question

pseudo-mhz .. are you a salesman?
By killing cpu, specifically meant burying it, no more functions.
I'll stay 533 , next to the smallest 533 cpu with 1mb (2.4a which should
have been labeled 2.4E, there IS a 2.26e nowhere to be found...)
I built a very silent, very cool setup,with rather large lightweight
aftermarket cpu heatsink. the 100w cooling values don't float my boat. It
could take it in OEM screaming fan glamour
but I don't want to. Yes , I agree some specific programs will use it more,
but as far as work's width, multitasking, that was my question.The increase
in memory bandwidth is a bonus going to 800. This 333mhz is so very close to
a 133 mhz bus I once ran with similar numbers. I could vomit.I may
reconsider for 800fsb, or, I may say to hell with the whole thing and run
with my complaints until there is truly a feat of engineering that will keep
me enthused.
thank you for reply.


"kony" wrote in message
...
On Mon, 13 Feb 2006 02:42:57 GMT, "bgd"
wrote:

Hi,
I buried my cpu, just tinkering to raise temp to my custom cooling. My
512kb l2 cahe p4 died at same spot as an intel p2 i had..(close to it).
Multithreading better than ever? Anyhow, my friend had the 1mb l2 cache p4
on same 533 bus, it did alot of work more before same choking.
Prioritizing
didn't seem to exist! (this is only after truly conquering heat to an
extreme) the only drawback is the cpu killed his mobo and was very hot
runner at loaded. Is the 800FSB better with 1mb l2 cache or does it
matter?
I like the 533 and only want increase of L2, but I don't know all the
facts,
and can't find an end users result.Any advice appreciated.


I assume the first couple sentences you wrote are important,
but they don't make a lot of sense within the context of the
next lines.

It would be good to get your cooling ironed out, what good
is custom cooling if you kill cpus? Plain old stock cooling
has to be better than that.

The benefit of a larger L2 cache depends on the task. Some
respond better than others and if you have a particular
need, certain tasks that stress the CPU most then those
would have to be considered. Generally speaking you would
want the 800 pseudo-MHz (since it's quad-pumped, 200MHz
clocked) for the best performance, or rather, having the
memory bus faster than 133MHz / DDR266 as well.



  #4  
Old February 13th 06, 09:38 AM posted to alt.comp.hardware
external usenet poster
 
Posts: n/a
Default L2 cache question

On Mon, 13 Feb 2006 04:38:26 GMT, "bgd"
wrote:

pseudo-mhz .. are you a salesman?


No but intel is. You see, it's not MHz by the industry
standards which use clock rate. The clock rate is not 800,
it's 200. Ever heard of DDR400? That "DDR" term is there
for a reason, in the case of the 800 pseudo-mhz, it should
be called QDR800, for quad data rate.


By killing cpu, specifically meant burying it, no more functions.


I have no idea what you mean. Bury with a shovel in the
garden?



I'll stay 533 , next to the smallest 533 cpu with 1mb (2.4a which should
have been labeled 2.4E, there IS a 2.26e nowhere to be found...)
I built a very silent, very cool setup,with rather large lightweight
aftermarket cpu heatsink. the 100w cooling values don't float my boat. It
could take it in OEM screaming fan glamour
but I don't want to.


I agree, loud heatsink fans are not so nice to use
long-term.


Yes , I agree some specific programs will use it more,
but as far as work's width, multitasking, that was my question.The increase
in memory bandwidth is a bonus going to 800. This 333mhz is so very close to
a 133 mhz bus I once ran with similar numbers. I could vomit.I may
reconsider for 800fsb, or, I may say to hell with the whole thing and run
with my complaints until there is truly a feat of engineering that will keep
me enthused.
thank you for reply.


Well one need not be enthused, only to realize that whatever
tasks they're doing either do, or do not, benefit enough
from an upgrade to make it worth the time or money.
  #5  
Old February 13th 06, 09:41 AM posted to alt.comp.hardware
external usenet poster
 
Posts: n/a
Default L2 cache question

In article 63UHf.3100$lG.1823@trndny01, "bgd" wrote:

pseudo-mhz .. are you a salesman?
By killing cpu, specifically meant burying it, no more functions.
I'll stay 533 , next to the smallest 533 cpu with 1mb (2.4a which should
have been labeled 2.4E, there IS a 2.26e nowhere to be found...)
I built a very silent, very cool setup,with rather large lightweight
aftermarket cpu heatsink. the 100w cooling values don't float my boat. It
could take it in OEM screaming fan glamour
but I don't want to. Yes , I agree some specific programs will use it more,
but as far as work's width, multitasking, that was my question.The increase
in memory bandwidth is a bonus going to 800. This 333mhz is so very close to
a 133 mhz bus I once ran with similar numbers. I could vomit.I may
reconsider for 800fsb, or, I may say to hell with the whole thing and run
with my complaints until there is truly a feat of engineering that will keep
me enthused.
thank you for reply.


I think the substance of this thread, is you are looking for a
computer with good performance, low heat, and low cost. I would say
pick two of three, because the pricing structures of things usually
means you will pay a lot more for the privilege of a cool running
system.

To a first order approximation, the first thing to optimize is
core clock rate. A 3.8GHz processor with a crippled cache structure
is still going to crush a 2.4Ghz processor with a slightly
nicer cache on it.

When you keep the core frequency constant, and then mess with
the front side bus or the cache, you are dealing with secondary
effects. FSBs up to a certain point now, are free, in the sense
that there is no reason for a price premium between FSB533 and
FSB800. Intel still charges a premium for FSB1066, but there
doesn't seem to be too much advantage going all the way to
FSB1066.

If you look at the AMD P.R. rating system, you can see that
doubling the cache size, maybe adds 200-300Mhz to the effective
computing rate. When you compare a Prescott to a Northwood,
the cache change cannot be compared directly, as the core engine
of the two processors is different - the larger cache was put
on the Prescott, not to give you twice the cache performance,
but to help to compensate for the characteristics of the
Prescott. You would need some careful benchmarks of the two
processors, to work out what the doubling is buying in that
case. Remember that some early benchmarks found the Prescott
slower for some things than the Northwood, and faster at
other things. So a direct comparison is difficult.

On socket 478, there are a couple of solutions with better
cooling performance. These solutions will not work optimally
well in any arbitrary S478 board, but they are solutions
nonetheless.

The first one is the Mobile Pentium4. If you take a mobile
chip (and there are several different kinds, so you have
to be extremely careful when shopping for these - I know,
because I got the wrong one), and put it in a desktop
motherboard, it runs 1200MHz core clock at FSB400. It doesn't
matter what the nominal speed marked in the device, they
all do the same thing (it is because they are using the
Speedstep "low gear").

If you stick one of those processors in a motherboard with
an 875P Northbridge, you can raise the FSB to FSB1200. The
core of the processor runs at 3.6GHz as a result. Even
low rated mobiles (like 1.8Ghz), can be raise to 3.6GHz.
There were people on Abxzone who were doing this stuff.
The processor stepping is important, and the D1 stepping
is the one required. The A and B steppings are not good
candidates, because they don't overclock all the way to
3.6GHz.

The primo processor for this, was the SL726. You can search
for that with your favorite search engine, or go to Abxzone
and read up on it there. The processor is supposed to run
cool, even at 3.6GHz. You can enter SL726 on
processorfinder.intel.com for more info.

http://processorfinder.intel.com/scr...sp?sSpec=SL726

A second option, is to use a Pentium-M processor in an
Asus S479 to S478 socket adapter (CT479). In this case,
the processors are expensive, but the operating power is
around 30W at stock speed. A limited number of Asus
motherboard models have modified BIOSes provided by
Asus to make this work.

As for your claims to killing the P4 processor. The
P4 family has a couple of protection features. When
the temperature goes over 70C, the processors can
throttle back the effective computing rate. The
clock inside the processor is gated off for short
intervales, which helps to control the temperature.

In the event that the heatsink falls off a P4 or an
Athlon64, I believe both platforms are now protected
by overheat detection. The latest Pentium4 processors
will turn off the computer when they hit 90C. Some
earlier processors turn off at 135C or so (the IC
package can take more heat, and that is why the temp
is set higher).

So, yes, you may have killed your processor somehow,
but I don't see how it can be temperature related.

Paul
 




Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

vB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Forum Jump

Similar Threads
Thread Thread Starter Forum Replies Last Post
P5A motherboard, socket 7 [email protected] Asus Motherboards 21 August 24th 04 05:02 PM
Strange beeps: Can L2 cache be faulty on Pentium III? Roger General 1 January 11th 04 03:26 PM
IBM white paper on Opteron Yousuf Khan General 115 November 7th 03 04:04 AM
Enabling write cache for a Maxtor DM9 160 GB PATA on a GigaRAID(ITE) or Sil3112? Robert Renner Storage (alternative) 0 September 28th 03 03:35 AM
New Hard Drive: 2MB cache vs. 8MB Cache Purp1e Storage (alternative) 9 September 18th 03 05:52 AM


All times are GMT +1. The time now is 02:15 AM.


Powered by vBulletin® Version 3.6.4
Copyright ©2000 - 2024, Jelsoft Enterprises Ltd.
Copyright ©2004-2024 HardwareBanter.
The comments are property of their posters.