Corsair memory wierdness with A7N8X's (way to save money)

G

Guest

Guest
Archived from groups: alt.comp.periphs.mainboard.asus (More info?)

I have two identical A7N8X systems, except that for one
I accidentally ordered a different grade of Corsair XMS
PC3200 memory. In one machine I have stuff that runs
at 200Mhz at 6-3-3-2, manually set since SPD is wrong.
In the other, I have a lower grade which runs at 8-3-3-3
at 200Mhz (and will fail if set lower). Both have the same
processor (3200+).

The weird thing is that these two systems perform
identically! In fact, I was at first getting lower numbers
on the 6-3-3-2 machine until I found that it had a debug
ethernet driver that was costing me a couple percent of
the CPU.

The symmetry in performance holds for all benchmarks,
including SiSoft Sandra memory bandwidth. Both get
2950/2750. So why pay extra for the lower latency
stuff if they perform the same? I saved about $50 due
to this apparently fortunate accident.

Maybe I am doing something wrong with the faster
memory, but I'm at a loss to figure it out. Is anyone
getting more than 2950 for the bandwidth with the
PC3200C2 at 200Mhz?

Alan
 
Archived from groups: alt.comp.periphs.mainboard.asus (More info?)

Corsair XMS
SiSoft Sandra says 3044 (no overclocking)


"Alan Wright" <alanatyahoodotcom> wrote in message
news:wPCdnXlQx9UmHR3dRVn-jg@giganews.com...
> I have two identical A7N8X systems, except that for one
> I accidentally ordered a different grade of Corsair XMS
> PC3200 memory. In one machine I have stuff that runs
> at 200Mhz at 6-3-3-2, manually set since SPD is wrong.
> In the other, I have a lower grade which runs at 8-3-3-3
> at 200Mhz (and will fail if set lower). Both have the same
> processor (3200+).
>
> The weird thing is that these two systems perform
> identically! In fact, I was at first getting lower numbers
> on the 6-3-3-2 machine until I found that it had a debug
> ethernet driver that was costing me a couple percent of
> the CPU.
>
> The symmetry in performance holds for all benchmarks,
> including SiSoft Sandra memory bandwidth. Both get
> 2950/2750. So why pay extra for the lower latency
> stuff if they perform the same? I saved about $50 due
> to this apparently fortunate accident.
>
> Maybe I am doing something wrong with the faster
> memory, but I'm at a loss to figure it out. Is anyone
> getting more than 2950 for the bandwidth with the
> PC3200C2 at 200Mhz?
>
> Alan
>
>
 
Archived from groups: alt.comp.periphs.mainboard.asus (More info?)

Timings?

Alan

"Mike Butler" <mnb0@flash.net> wrote in message
news:Oy%fc.17935$9P6.13614@newssvr22.news.prodigy.com...
>
> Corsair XMS
> SiSoft Sandra says 3044 (no overclocking)
>
>
> "Alan Wright" <alanatyahoodotcom> wrote in message
> news:wPCdnXlQx9UmHR3dRVn-jg@giganews.com...
> > I have two identical A7N8X systems, except that for one
> > I accidentally ordered a different grade of Corsair XMS
> > PC3200 memory. In one machine I have stuff that runs
> > at 200Mhz at 6-3-3-2, manually set since SPD is wrong.
> > In the other, I have a lower grade which runs at 8-3-3-3
> > at 200Mhz (and will fail if set lower). Both have the same
> > processor (3200+).
> >
> > The weird thing is that these two systems perform
> > identically! In fact, I was at first getting lower numbers
> > on the 6-3-3-2 machine until I found that it had a debug
> > ethernet driver that was costing me a couple percent of
> > the CPU.
> >
> > The symmetry in performance holds for all benchmarks,
> > including SiSoft Sandra memory bandwidth. Both get
> > 2950/2750. So why pay extra for the lower latency
> > stuff if they perform the same? I saved about $50 due
> > to this apparently fortunate accident.
> >
> > Maybe I am doing something wrong with the faster
> > memory, but I'm at a loss to figure it out. Is anyone
> > getting more than 2950 for the bandwidth with the
> > PC3200C2 at 200Mhz?
> >
> > Alan
> >
> >
>
>
 
Archived from groups: alt.comp.periphs.mainboard.asus (More info?)

>I have two identical A7N8X systems, except that for one
>I accidentally ordered a different grade of Corsair XMS
>PC3200 memory. In one machine I have stuff that runs
>at 200Mhz at 6-3-3-2, manually set since SPD is wrong.
>In the other, I have a lower grade which runs at 8-3-3-3
>at 200Mhz (and will fail if set lower). Both have the same
>processor (3200+).
>
>The weird thing is that these two systems perform
>identically! In fact, I was at first getting lower numbers
>on the 6-3-3-2 machine until I found that it had a debug
>ethernet driver that was costing me a couple percent of
>the CPU.
>
>The symmetry in performance holds for all benchmarks,
>including SiSoft Sandra memory bandwidth. Both get
>2950/2750. So why pay extra for the lower latency
>stuff if they perform the same? I saved about $50 due
>to this apparently fortunate accident.
>
>Maybe I am doing something wrong with the faster
>memory, but I'm at a loss to figure it out. Is anyone
>getting more than 2950 for the bandwidth with the
>PC3200C2 at 200Mhz?
>
>Alan
>
For a better RAM score change that 6 to 10 or 11.
 
Archived from groups: alt.comp.periphs.mainboard.asus (More info?)

"Andrew J" <ajpk3@hotmail.comremove> wrote in message
news😱s61805bhq0si1f0966tees4l4kuro3inv@4ax.com...
>
>
> >I have two identical A7N8X systems, except that for one
> >I accidentally ordered a different grade of Corsair XMS
> >PC3200 memory. In one machine I have stuff that runs
> >at 200Mhz at 6-3-3-2, manually set since SPD is wrong.
> >In the other, I have a lower grade which runs at 8-3-3-3
> >at 200Mhz (and will fail if set lower). Both have the same
> >processor (3200+).
> >
> >The weird thing is that these two systems perform
> >identically! In fact, I was at first getting lower numbers
> >on the 6-3-3-2 machine until I found that it had a debug
> >ethernet driver that was costing me a couple percent of
> >the CPU.
> >
> >The symmetry in performance holds for all benchmarks,
> >including SiSoft Sandra memory bandwidth. Both get
> >2950/2750. So why pay extra for the lower latency
> >stuff if they perform the same? I saved about $50 due
> >to this apparently fortunate accident.
> >
> >Maybe I am doing something wrong with the faster
> >memory, but I'm at a loss to figure it out. Is anyone
> >getting more than 2950 for the bandwidth with the
> >PC3200C2 at 200Mhz?
> >
> >Alan
> >
> For a better RAM score change that 6 to 10 or 11.

Why is that?

I did try changing it to 8-3-3-3 to match the other system.
There was a small decrease in scores. Maybe noise.

Alan
 
Archived from groups: alt.comp.periphs.mainboard.asus (More info?)

On Fri, 16 Apr 2004 18:56:30 -0600, "Alan Wright" <alanatyahoodotcom> wrote:

>I have two identical A7N8X systems, except that for one
>I accidentally ordered a different grade of Corsair XMS
>PC3200 memory. In one machine I have stuff that runs
>at 200Mhz at 6-3-3-2, manually set since SPD is wrong.
>In the other, I have a lower grade which runs at 8-3-3-3
>at 200Mhz (and will fail if set lower). Both have the same
>processor (3200+).
>
>The weird thing is that these two systems perform
>identically! In fact, I was at first getting lower numbers
>on the 6-3-3-2 machine until I found that it had a debug
>ethernet driver that was costing me a couple percent of
>the CPU.
>
>The symmetry in performance holds for all benchmarks,
>including SiSoft Sandra memory bandwidth. Both get
>2950/2750. So why pay extra for the lower latency
>stuff if they perform the same? I saved about $50 due
>to this apparently fortunate accident.
>
>Maybe I am doing something wrong with the faster
>memory, but I'm at a loss to figure it out. Is anyone
>getting more than 2950 for the bandwidth with the
>PC3200C2 at 200Mhz?
>
>Alan
>

Have your tried it with buffering off in sandra?
some of my unbuffered tests - http://tinyurl.com/38ozj

Ed