|
|
|
|
| Author |
Message |
Thomas
Guest
|
 Posted: Sun Apr 29, 2007 12:29 am Post subject: Very meagre E4300 results... snif |
 |
|
Today I finally picked up my complete system. I got lucky on the case, they
had a 'slightly damaged' Antec Performance One for half the price. The
damage (invisible) didnt bothe rme so I jumped on the opportunity. I added
an Antec True Power Trio 650, so I at least wouldn't have to worry about
that end of my purchase :-)
Anyway, I built the whole thing, and for the first time in my life, the new
PC worked from scratch :-)
A short list:
- Antec Performace One P180 + Antec Truepower Trio 650W
- Seagate 320 GB 7200 rpm 16 MB
- Samsung 18x DVDRW
- Asus P5N-E SLI
- Kingston 2x 1 GB PC6400-LL
- EVGA 8800GTS 320
- Intel E4300
- Arctic Cooling Freezer Pro 7 cooler
When I first booted into the BIOS, I was amazed at all the little things I
could tweak. Even compared to my P4C from three years ago, it's very
impressive.
Anyway, I set the FSB to 1000 MHz ans started my Windows install. After
that, I overclocked a bit more methodically :-)
Memory settings: 5 / 5 / 5 / 5 / 13 - 2T / 800 MHz (conservative settings)
FSB unlocked from memory clock.
In increments of 10 MHz I raised my FSB:
1.38 V worked until 1090 MHz
1.425 V worked until 1120 MHz
1.45 V worked until 1150 MHz
I did an Orthos burn in at 1066 MHz / 1.40 volts. This resulted in a
stressed temp of 43°C (34 idle)
Frankly, I'm very unimpressed... Can anyone guess at the reason my OC is
this low? I'm not calling quits yet, but it's not a very satisfying
experience :-)
--
Met vriendelijke groeten, Thomas van der Horst |
|
| Back to top |
|
 |
|
|
Thomas
Guest
|
| Posted: Sun Apr 29, 2007 2:10 pm Post subject: Re: Very meagre E4300 results... snif |
|
|
Phil Weldon wrote:
| Quote: | In increments of 10 MHz I raised my FSB:
1.38 V worked until 1090 MHz
1.425 V worked until 1120 MHz
1.45 V worked until 1150 MHz
I did an Orthos burn in at 1066 MHz / 1.40 volts. This resulted in a
stressed temp of 43°C (34 idle)
Frankly, I'm very unimpressed... Can anyone guess at the reason my
OC is this low? I'm not calling quits yet, but it's not a very
satisfying experience :-)
_____
I have had very poor results when I try settings through nTune and
very unreliable temperature / voltage / fan speed reports through
nMontior and nTune.
I will download the manual for your ASUS motherboard and compare the
labels ASUS uses for parameters in the BIOS to the labels EVGA uses.
Perhaps then I can post further suggestions. Of course, I have to
get the ASUS download site to actually WORK. Not being able to
download ASUS manuals is the main reason I did not go with an ASUS
motherboard!
You could try backing down the CPU core voltage checking for small
overclocks. This MIGHT give you useful information. Also I think I
have my chipset voltage increased by 0.1 volts.
|
Thanks for the heads-up. I've raised NB voltage to 1.4 Volts, but this
doesnt help the OC at all. I've found a nice PDF from NVidia
(nforce_680i_sli_overclocking.pdf) which explains overclocking on the 680i
very specifically, I'll use that to see if I did something wrong.
I'm uploading both the P5N-E SLI manual and the overclocking document to the
following location:
http://www.eastwestjewels.nl/temp/
Thanks for your help, Phil!
By the way, I'm more lucky with my 8800GTS; I can OC it from 500/800 stock
to 620/950! It's just getting too hot at those clocks, so for games I've
selected 600/900.
--
Met vriendelijke groeten, Thomas van der Horst |
|
| Back to top |
|
|
Jack R
Guest
|
| Posted: Sun Apr 29, 2007 2:51 pm Post subject: Re: Very meagre E4300 results... snif |
|
|
--stuff deleted--
| Quote: | I will download the manual for your ASUS motherboard and compare the
labels
ASUS uses for parameters in the BIOS to the labels EVGA uses. Perhaps
then
I can post further suggestions. Of course, I have to get the ASUS
download
site to actually WORK. Not being able to download ASUS manuals is the
main
reason I did not go with an ASUS motherboard!
--more stuff deleted-- |
Try downloading from ftp.asus.com.tw
I've had good luck with this, including later drivers than can be found on
their web site.
Jack R |
|
| Back to top |
|
|
Phil Weldon
Guest
|
| Posted: Mon Apr 30, 2007 12:50 am Post subject: Re: Very meagre E4300 results... snif |
|
|
Thanks for the ASUS P5N-E SLI manual upload.
The EVGA 680i BIOS does not have the 'SLI Broadcast Aperture' entry (which I
don't think could have any effect on overclocking the CPU.
The EVGA has what I think is the 'LDT Frequency' for the ASUS P5N-E SLI
BIOS. I use [5X].
The bus voltages are set to automatic. That gives an FSB voltage of 1.4000
v and a PCI-E voltage of 1.45 v. The default voltages are set for
HT(SPP-MPC) [1.2000 v], nForce [1.5000 v], and AUX [1.5000].
I also have all 'Spread Spectrum' settings at [Disabled].
How useful is ASUS Probe?
I think getting reliable temperature readings is a problem.
When my system is running at 2.7 GHz (FSB 1200 MHz / Memory bus 1200 MHz /
core 1.250 v / memory 2.3 v / Intel boxed retail) with Orthos (small FFTs)
the 35 C (23 C ambient) temperature rapidly rises through 67 C. At about
that temperature point Orthos throws an error. When I am lucky enough to
have a 22 C or lower ambient Orthos will continue indefinitely. The other
Orthos settings stress the CPUs much less (the temperatures level out at 61
C to 65 C. The above temperatures are all reported by nMonitor.
Occasionaly Orthos will cause an error and be closed by Windows XP Pro.
Speed Fan reports CPU temperature, Core 0 temperature and Core 1
temperature. CPU temperature as reported by Speed Fan is 10 C higher than
the reported temperature for Core 0 and Core 1. Speed Fan also reports a
'System' temperature of - 65 C and an 'AUX' temperature of 38 C. Speed Fan
reports a core voltage of 1.2 volts. Speed fan CPU temperature can vary as
much as 3 C from the nMonitor CPU temperature, though usually the reported
CPU temperatures are the same or 1 C off.
*****Under the same overclocking conditions of
E4400 / EVGA 680i / 2.7 GHz / 1200 MHz FSB 1200 MHz Memory Bus / 1.250
core voltage / 23 C ambient
and
Orthos small FFTs
*****
Core Temp 0.95 reports
Tjunction: 100 C
Core #0: 72 C
Core #1: 72 C
*****
Speed Fab reports
CPU: 67 C
Core 0: 57C
Core 1: 57C
*****
nMonitor reports
CPU: 66 C
Phil Weldon
"Thomas" <ThomasH@lycosmail.nl> wrote in message
news:4634610a$0$16940$9a622dc7@news.kpnplanet.nl...
| Phil Weldon wrote:
| >> In increments of 10 MHz I raised my FSB:
| >> 1.38 V worked until 1090 MHz
| >> 1.425 V worked until 1120 MHz
| >> 1.45 V worked until 1150 MHz
| >>
| >> I did an Orthos burn in at 1066 MHz / 1.40 volts. This resulted in a
| >> stressed temp of 43°C (34 idle)
| >>
| >> Frankly, I'm very unimpressed... Can anyone guess at the reason my
| >> OC is this low? I'm not calling quits yet, but it's not a very
| >> satisfying experience :-)
| > _____
| >
| > I have had very poor results when I try settings through nTune and
| > very unreliable temperature / voltage / fan speed reports through
| > nMontior and nTune.
| >
| > I will download the manual for your ASUS motherboard and compare the
| > labels ASUS uses for parameters in the BIOS to the labels EVGA uses.
| > Perhaps then I can post further suggestions. Of course, I have to
| > get the ASUS download site to actually WORK. Not being able to
| > download ASUS manuals is the main reason I did not go with an ASUS
| > motherboard!
| >
| > You could try backing down the CPU core voltage checking for small
| > overclocks. This MIGHT give you useful information. Also I think I
| > have my chipset voltage increased by 0.1 volts.
|
| Thanks for the heads-up. I've raised NB voltage to 1.4 Volts, but this
| doesnt help the OC at all. I've found a nice PDF from NVidia
| (nforce_680i_sli_overclocking.pdf) which explains overclocking on the 680i
| very specifically, I'll use that to see if I did something wrong.
|
| I'm uploading both the P5N-E SLI manual and the overclocking document to
the
| following location:
| http://www.eastwestjewels.nl/temp/
|
| Thanks for your help, Phil!
|
| By the way, I'm more lucky with my 8800GTS; I can OC it from 500/800 stock
| to 620/950! It's just getting too hot at those clocks, so for games I've
| selected 600/900.
|
| --
| Met vriendelijke groeten, Thomas van der Horst
|
| |
|
| Back to top |
|
|
Fishface
Guest
|
| Posted: Mon Apr 30, 2007 6:37 am Post subject: Re: Very meagre E4300 results... snif |
|
|
Phil Weldon" <not.disclosed@example.com> wrote in message news:p39Zh.2192$296.1644@newsread4.news.pas.earthlink.net...
| Quote: | Thanks for the ASUS P5N-E SLI manual upload.
The EVGA 680i BIOS does not have the 'SLI Broadcast Aperture' entry (which I
don't think could have any effect on overclocking the CPU.
The EVGA has what I think is the 'LDT Frequency' for the ASUS P5N-E SLI
BIOS. I use [5X].
The bus voltages are set to automatic. That gives an FSB voltage of 1.4000
v and a PCI-E voltage of 1.45 v. The default voltages are set for
HT(SPP-MPC) [1.2000 v], nForce [1.5000 v], and AUX [1.5000].
I also have all 'Spread Spectrum' settings at [Disabled].
How useful is ASUS Probe?
I think getting reliable temperature readings is a problem.
When my system is running at 2.7 GHz (FSB 1200 MHz / Memory bus 1200 MHz /
core 1.250 v / memory 2.3 v / Intel boxed retail) with Orthos (small FFTs)
the 35 C (23 C ambient) temperature rapidly rises through 67 C. At about
that temperature point Orthos throws an error. When I am lucky enough to
have a 22 C or lower ambient Orthos will continue indefinitely. The other
Orthos settings stress the CPUs much less (the temperatures level out at 61
C to 65 C. The above temperatures are all reported by nMonitor.
Occasionaly Orthos will cause an error and be closed by Windows XP Pro.
Speed Fan reports CPU temperature, Core 0 temperature and Core 1
temperature. CPU temperature as reported by Speed Fan is 10 C higher than
the reported temperature for Core 0 and Core 1. Speed Fan also reports a
'System' temperature of - 65 C and an 'AUX' temperature of 38 C. Speed Fan
reports a core voltage of 1.2 volts. Speed fan CPU temperature can vary as
much as 3 C from the nMonitor CPU temperature, though usually the reported
CPU temperatures are the same or 1 C off.
*****Under the same overclocking conditions of
E4400 / EVGA 680i / 2.7 GHz / 1200 MHz FSB 1200 MHz Memory Bus / 1.250
core voltage / 23 C ambient
and
Orthos small FFTs
*****
Core Temp 0.95 reports
Tjunction: 100 C
Core #0: 72 C
Core #1: 72 C
*****
Speed Fab reports
CPU: 67 C
Core 0: 57C
Core 1: 57C
*****
nMonitor reports
CPU: 66 C
|
Well, that's quite a large range. I'd say every possible temperature is
about covered! Did you read the page about how CoreTemp works?
http://www.alcpu.com/CoreTemp/howitworks.html
I see your TJunction is "defined by Intel" to be 100° C. By comparison,
mine on the E6400 says 85° C. Although I wasn't using CoreTemp
when I had my retail HSF installed, I think those temperatures seem
very high for such a low voltage and mild overclock. Two possibilities
comes to mind with regard to this. First is the possible use of the
previously mentioned new practice of using thermal paste instead
of thermal epoxy under the IHS. The second would be a poor thermal
interface between the IHS and retail cooler. I examined my retail
cooler by holding my beveled machinist's square up to the smooth and
shiny contact surface of the cooler, which appears machined by turning.
To my dismay, the center was much lower than the edge, making it
effectively concave. This would not be easy to lap as the mounting
bracket does not seem to be removable and extends below the copper.
So now I'm looking to cut a thin piece of glass for use in lapping, and
hopefully, it will not cut me back!
Here's a thread about the IHS:
http://www.xtremesystems.org/forums/showthread.php?t=136776
Under the fan, I was surprised to see the copper is hollowed-out, as
seen in the picture on this page:
http://www.frostytech.com/articleview.cfm?articleID=2025
My fan turns slower than the one in the article, so it must be a slightly
different model from those used on the Pentium D. |
|
| Back to top |
|
|
Fix your Windows Problems - FAST.
FREE Safe Scan Registry Check. Locate & Fix Errors in Minutes!
|
|
Jack R
Guest
|
| Posted: Mon Apr 30, 2007 7:31 am Post subject: Re: Very meagre E4300 results... snif |
|
|
"Phil Weldon" <not.disclosed@example.com> wrote in message
news:p39Zh.2192$296.1644@newsread4.news.pas.earthlink.net...
| Quote: | Thanks for the ASUS P5N-E SLI manual upload.
The EVGA 680i BIOS does not have the 'SLI Broadcast Aperture' entry (which
I
don't think could have any effect on overclocking the CPU.
The EVGA has what I think is the 'LDT Frequency' for the ASUS P5N-E SLI
BIOS. I use [5X].
The bus voltages are set to automatic. That gives an FSB voltage of
1.4000
v and a PCI-E voltage of 1.45 v. The default voltages are set for
HT(SPP-MPC) [1.2000 v], nForce [1.5000 v], and AUX [1.5000].
I also have all 'Spread Spectrum' settings at [Disabled].
How useful is ASUS Probe?
I think getting reliable temperature readings is a problem.
When my system is running at 2.7 GHz (FSB 1200 MHz / Memory bus 1200 MHz /
core 1.250 v / memory 2.3 v / Intel boxed retail) with Orthos (small
FFTs)
the 35 C (23 C ambient) temperature rapidly rises through 67 C. At about
that temperature point Orthos throws an error. When I am lucky enough to
have a 22 C or lower ambient Orthos will continue indefinitely. The other
Orthos settings stress the CPUs much less (the temperatures level out at
61
C to 65 C. The above temperatures are all reported by nMonitor.
Occasionaly Orthos will cause an error and be closed by Windows XP Pro.
Speed Fan reports CPU temperature, Core 0 temperature and Core 1
temperature. CPU temperature as reported by Speed Fan is 10 C higher than
the reported temperature for Core 0 and Core 1. Speed Fan also reports a
'System' temperature of - 65 C and an 'AUX' temperature of 38 C. Speed
Fan
reports a core voltage of 1.2 volts. Speed fan CPU temperature can vary
as
much as 3 C from the nMonitor CPU temperature, though usually the reported
CPU temperatures are the same or 1 C off.
*****Under the same overclocking conditions of
E4400 / EVGA 680i / 2.7 GHz / 1200 MHz FSB 1200 MHz Memory Bus / 1.250
core voltage / 23 C ambient
and
Orthos small FFTs
*****
Core Temp 0.95 reports
Tjunction: 100 C
Core #0: 72 C
Core #1: 72 C
*****
Speed Fab reports
CPU: 67 C
Core 0: 57C
Core 1: 57C
*****
nMonitor reports
CPU: 66 C
Phil Weldon
"Thomas" <ThomasH@lycosmail.nl> wrote in message
news:4634610a$0$16940$9a622dc7@news.kpnplanet.nl...
| Phil Weldon wrote:
| >> In increments of 10 MHz I raised my FSB:
| >> 1.38 V worked until 1090 MHz
| >> 1.425 V worked until 1120 MHz
| >> 1.45 V worked until 1150 MHz
|
| >> I did an Orthos burn in at 1066 MHz / 1.40 volts. This resulted in a
| >> stressed temp of 43°C (34 idle)
|
| >> Frankly, I'm very unimpressed... Can anyone guess at the reason my
| >> OC is this low? I'm not calling quits yet, but it's not a very
| >> satisfying experience :-)
| > _____
|
| > I have had very poor results when I try settings through nTune and
| > very unreliable temperature / voltage / fan speed reports through
| > nMontior and nTune.
|
| > I will download the manual for your ASUS motherboard and compare the
| > labels ASUS uses for parameters in the BIOS to the labels EVGA uses.
| > Perhaps then I can post further suggestions. Of course, I have to
| > get the ASUS download site to actually WORK. Not being able to
| > download ASUS manuals is the main reason I did not go with an ASUS
| > motherboard!
|
| > You could try backing down the CPU core voltage checking for small
| > overclocks. This MIGHT give you useful information. Also I think I
| > have my chipset voltage increased by 0.1 volts.
|
| Thanks for the heads-up. I've raised NB voltage to 1.4 Volts, but this
| doesnt help the OC at all. I've found a nice PDF from NVidia
| (nforce_680i_sli_overclocking.pdf) which explains overclocking on the
680i
| very specifically, I'll use that to see if I did something wrong.
|
| I'm uploading both the P5N-E SLI manual and the overclocking document to
the
| following location:
| http://www.eastwestjewels.nl/temp/
|
| Thanks for your help, Phil!
|
| By the way, I'm more lucky with my 8800GTS; I can OC it from 500/800
stock
| to 620/950! It's just getting too hot at those clocks, so for games I've
| selected 600/900.
|
| --
| Met vriendelijke groeten, Thomas van der Horst
|
|
On my P5N-E Sli, Core Temp 0.95 gives the same cpu temps as SpeedFan 4.32, |
within 1 deg C, even in dynamic conditions. ASUS Probe II lags these
measurements by _minutes_ at best.
What guide can be trusted for system voltages? Everything I've seen varies
all over the place.
Thanks,
Jack R
(Unable to get my E6600 stable above 3.150GHz) |
|
| Back to top |
|
|
Thomas
Guest
|
| Posted: Mon Apr 30, 2007 11:05 am Post subject: Re: Very meagre E4300 results... snif |
|
|
Phil Weldon wrote:
| Quote: | Thanks for the ASUS P5N-E SLI manual upload.
The EVGA 680i BIOS does not have the 'SLI Broadcast Aperture' entry
(which I don't think could have any effect on overclocking the CPU.
The EVGA has what I think is the 'LDT Frequency' for the ASUS P5N-E
SLI BIOS. I use [5X].
The bus voltages are set to automatic. That gives an FSB voltage of
1.4000 v and a PCI-E voltage of 1.45 v. The default voltages are set
for HT(SPP-MPC) [1.2000 v], nForce [1.5000 v], and AUX [1.5000].
|
You're welcome, of course... Anyway, I can only set VCore (1.325 - 1.45
tested), northbridge (1.2 - 1.4 tested), memory (2.0 - 2.2 tested) and VCore
offset (didnt use).
| Quote: | I also have all 'Spread Spectrum' settings at [Disabled].
|
Ditto.
| Quote: | How useful is ASUS Probe?
|
Didnt try it, as I've had bad experiences in the past (unreliable, unstable,
heavy resources...)
| Quote: | I think getting reliable temperature readings is a problem.
When my system is running at 2.7 GHz (FSB 1200 MHz / Memory bus 1200
MHz / core 1.250 v / memory 2.3 v / Intel boxed retail) with Orthos
(small FFTs) the 35 C (23 C ambient) temperature rapidly rises
through 67 C. At about that temperature point Orthos throws an
error. When I am lucky enough to have a 22 C or lower ambient Orthos
will continue indefinitely. The other Orthos settings stress the
CPUs much less (the temperatures level out at 61 C to 65 C. The
above temperatures are all reported by nMonitor. Occasionaly Orthos
will cause an error and be closed by Windows XP Pro.
Speed Fan reports CPU temperature, Core 0 temperature and Core 1
temperature. CPU temperature as reported by Speed Fan is 10 C higher
than the reported temperature for Core 0 and Core 1. Speed Fan also
reports a 'System' temperature of - 65 C and an 'AUX' temperature of
38 C. Speed Fan reports a core voltage of 1.2 volts. Speed fan CPU
temperature can vary as much as 3 C from the nMonitor CPU
temperature, though usually the reported CPU temperatures are the
same or 1 C off.
snipped many readings |
Speedfan and CoreTemp get the same readings for the core temperatures.
There's a huge discrepancy between these temps and the 'CPU temp', though.
At idle, there's a gap of about 10 °C between the two, while at stressed,
the temps are roughly the same. See these screenshots:
http://www.eastwestjewels.nl/temp/stressed.PNG
http://www.eastwestjewels.nl/temp/idle.PNG
Right now, I'm running at 1066 MHz, RAM at 533, very low timings
(4/3/3/13/1T). At these timings, SANDRA performance is 5% less than with the
memory @ 800, 4/4/4/13/1T timings.
--
Met vriendelijke groeten, Thomas van der Horst |
|
| Back to top |
|
|
Ed Medlin
Guest
|
| Posted: Mon Apr 30, 2007 4:06 pm Post subject: Re: Very meagre E4300 results... snif |
|
|
"Thomas" <ThomasH@lycosmail.nl> wrote in message
news:4635872c$0$16946$9a622dc7@news.kpnplanet.nl...
| Quote: | Phil Weldon wrote:
Thanks for the ASUS P5N-E SLI manual upload.
The EVGA 680i BIOS does not have the 'SLI Broadcast Aperture' entry
(which I don't think could have any effect on overclocking the CPU.
The EVGA has what I think is the 'LDT Frequency' for the ASUS P5N-E
SLI BIOS. I use [5X].
The bus voltages are set to automatic. That gives an FSB voltage of
1.4000 v and a PCI-E voltage of 1.45 v. The default voltages are set
for HT(SPP-MPC) [1.2000 v], nForce [1.5000 v], and AUX [1.5000].
You're welcome, of course... Anyway, I can only set VCore (1.325 -
1.45 tested), northbridge (1.2 - 1.4 tested), memory (2.0 - 2.2
tested) and VCore offset (didnt use).
I also have all 'Spread Spectrum' settings at [Disabled].
Ditto.
How useful is ASUS Probe?
Didnt try it, as I've had bad experiences in the past (unreliable,
unstable, heavy resources...)
I think getting reliable temperature readings is a problem.
When my system is running at 2.7 GHz (FSB 1200 MHz / Memory bus 1200
MHz / core 1.250 v / memory 2.3 v / Intel boxed retail) with Orthos
(small FFTs) the 35 C (23 C ambient) temperature rapidly rises
through 67 C. At about that temperature point Orthos throws an
error. When I am lucky enough to have a 22 C or lower ambient Orthos
will continue indefinitely. The other Orthos settings stress the
CPUs much less (the temperatures level out at 61 C to 65 C. The
above temperatures are all reported by nMonitor. Occasionaly Orthos
will cause an error and be closed by Windows XP Pro.
Speed Fan reports CPU temperature, Core 0 temperature and Core 1
temperature. CPU temperature as reported by Speed Fan is 10 C higher
than the reported temperature for Core 0 and Core 1. Speed Fan also
reports a 'System' temperature of - 65 C and an 'AUX' temperature of
38 C. Speed Fan reports a core voltage of 1.2 volts. Speed fan CPU
temperature can vary as much as 3 C from the nMonitor CPU
temperature, though usually the reported CPU temperatures are the
same or 1 C off.
snipped many readings
Speedfan and CoreTemp get the same readings for the core temperatures.
There's a huge discrepancy between these temps and the 'CPU temp',
though. At idle, there's a gap of about 10 °C between the two, while
at stressed, the temps are roughly the same. See these screenshots:
http://www.eastwestjewels.nl/temp/stressed.PNG
http://www.eastwestjewels.nl/temp/idle.PNG
Right now, I'm running at 1066 MHz, RAM at 533, very low timings
(4/3/3/13/1T). At these timings, SANDRA performance is 5% less than
with the memory @ 800, 4/4/4/13/1T timings.
Funny how that works isn't it? I have found the same results before in a |
few cases. Tighter timings don't always give better results even though
it is stable. There seems to be a 'sweet spot' with most memory when it
will OC well and also be stable and bench test faster. I don't have the
answer, just an observation.......:-).
Ed |
|
| Back to top |
|
|
|
|
| |
Watched Topics
Search
Register
Log in to check your private messages
Profile
Log in
Memberlist
Usergroups