Quad core benching like a dual

  • Thread starter Thread starter Power Obsessed
  • Start date Start date
Re: Quad core benching like a dual

Most have multiple layers of sensitivity. Say, one speed at 25%, another at
40% and above say 66% full speed. You should be able to check Intel's site
for where their models adjust.

"Power Obsessed" <screenedemail@comcast.net> wrote in message
news:0D6ABD17-7C3A-4D27-BAA3-51D1AC33008D@microsoft.com...
> Ya but that only happens when its idle right? Not in the middle of the
> processor being called upon. I know it will automatically shut the system
> down if you reach critical temps.
> --
> Asus P5W DH , Q6600 Quad core 2.4ghz @3.12, thermalake V1 heatsink, 4gb
> Crucial ballistix 800mhz , 16gb Mtron 6000 SSD, 150gb Raptor X, Nvidia
> 8800gt , 700 watt OCZ GameXStream power s , XP Pro X64
>
>
> "Colin Barnhorst" wrote:
>
>> Modern processors throttle down to conserve energy and improve thermal
>> conditions.
>>
>> "Power Obsessed" <screenedemail@comcast.net> wrote in message
>> news:88425497-9428-4AC6-AC76-0A82D2FD275E@microsoft.com...
>> > To Zootal:
>> > The thing that really threw me was the HUGE difference in the results.
>> > (from 1600+ to 800+). I had used this software in the past "right out
>> > of
>> > the
>> > box" I guess you could say. No configuration at all.
>> > The setting I'm talking about is the number of processes to test the
>> > CPU
>> > with. I have no idea why that setting was different
>> > Thank,
>> > Jim
>> > --
>> > Asus P5W DH , Q6600 Quad core 2.4ghz @3.12, thermalake V1 heatsink, 4gb
>> > Crucial ballistix 800mhz , 16gb Mtron 6000 SSD, 150gb Raptor X,
>> > Nvidia
>> > 8800gt , 700 watt OCZ GameXStream power s , XP Pro X64
>> >
>> >
>> > "Zootal" wrote:
>> >
>> >> In addition to all of the varied comments that others have added....I
>> >> might
>> >> suggest that you not put a lot of emphasis or confidence on any one
>> >> benchmark. Processors vary greatly internally, and so do benchmarks.
>> >> It's
>> >> not difficult to tune a benchmark to favor a specific architecture or
>> >> configuration. This happens, and not always intentionally. Anytime we
>> >> change
>> >> something in a compiler or cpu config, etc., we run a half dozen or so
>> >> benchmarks so that we get a better profiling of what the change did,
>> >> and
>> >> expose potential problems. No one benchmark can tell it all, and none
>> >> of
>> >> them are representative of what you will actually be doing with your
>> >> computer.
>> >>
>> >>
>> >>
>> >>
>> >> <miso@sushi.com> wrote in message
>> >> news:e4934900-f01b-4282-a008-829053bb7823@59g2000hsb.googlegroups.com...
>> >> > On Mar 28, 8:36 pm, Power Obsessed <screenedem...@comcast.net>
>> >> > wrote:
>> >> >> I ended up figuring it out. If your interested it was the way the
>> >> >> benchmark
>> >> >> software was configured. I do want to thank all of you for being so
>> >> >> helpful.
>> >> >>
>> >> >> Jim
>> >>
>> >>
>> >>
>>
 
Re: Quad core benching like a dual

Yeah, that would definitely be significant :D:D:D

"Power Obsessed" <screenedemail@comcast.net> wrote in message
news:88425497-9428-4AC6-AC76-0A82D2FD275E@microsoft.com...
> To Zootal:
> The thing that really threw me was the HUGE difference in the results.
> (from 1600+ to 800+). I had used this software in the past "right out of
> the
> box" I guess you could say. No configuration at all.
> The setting I'm talking about is the number of processes to test the CPU
> with. I have no idea why that setting was different
> Thank,
> Jim
> --
> Asus P5W DH , Q6600 Quad core 2.4ghz @3.12, thermalake V1 heatsink, 4gb
> Crucial ballistix 800mhz , 16gb Mtron 6000 SSD, 150gb Raptor X, Nvidia
> 8800gt , 700 watt OCZ GameXStream power s , XP Pro X64
>
>
> "Zootal" wrote:
>
>> In addition to all of the varied comments that others have added....I
>> might
>> suggest that you not put a lot of emphasis or confidence on any one
>> benchmark. Processors vary greatly internally, and so do benchmarks. It's
>> not difficult to tune a benchmark to favor a specific architecture or
>> configuration. This happens, and not always intentionally. Anytime we
>> change
>> something in a compiler or cpu config, etc., we run a half dozen or so
>> benchmarks so that we get a better profiling of what the change did, and
>> expose potential problems. No one benchmark can tell it all, and none of
>> them are representative of what you will actually be doing with your
>> computer.
>>
>>
>>
>>
>> <miso@sushi.com> wrote in message
>> news:e4934900-f01b-4282-a008-829053bb7823@59g2000hsb.googlegroups.com...
>> > On Mar 28, 8:36 pm, Power Obsessed <screenedem...@comcast.net> wrote:
>> >> I ended up figuring it out. If your interested it was the way the
>> >> benchmark
>> >> software was configured. I do want to thank all of you for being so
>> >> helpful.
>> >>
>> >> Jim
>>
>>
>>
 
Re: Quad core benching like a dual

Desktop processors (pentiums, core, athlon, phenom, etc.) won't do it under
load unless they are overheating. Lowering the voltage to unused functional
units is an old technique, but they only do it when they think it's not
needed. Specialty processors, OTOH, will do it even under load, depending on
what it's currently doing. If we can scale the voltage down and still get
the work done on time, then we do it. If I can distribute the load among two
cores and lower the voltage and clock, I get the work done in the same
amount of time, but use half the power. I don't need balls-to-the-wall
processing 95% of the time.

"Colin Barnhorst" <c.barnhorst@comcast.net> wrote in message
news:44CD7C1B-C8D8-4E31-8D7C-3BC7412942B2@microsoft.com...
> Modern processors throttle down to conserve energy and improve thermal
> conditions.
>
> "Power Obsessed" <screenedemail@comcast.net> wrote in message
> news:88425497-9428-4AC6-AC76-0A82D2FD275E@microsoft.com...
>> To Zootal:
>> The thing that really threw me was the HUGE difference in the results.
>> (from 1600+ to 800+). I had used this software in the past "right out of
>> the
>> box" I guess you could say. No configuration at all.
>> The setting I'm talking about is the number of processes to test the CPU
>> with. I have no idea why that setting was different
>> Thank,
>> Jim
>> --
>> Asus P5W DH , Q6600 Quad core 2.4ghz @3.12, thermalake V1 heatsink, 4gb
>> Crucial ballistix 800mhz , 16gb Mtron 6000 SSD, 150gb Raptor X, Nvidia
>> 8800gt , 700 watt OCZ GameXStream power s , XP Pro X64
>>
>>
>> "Zootal" wrote:
>>
>>> In addition to all of the varied comments that others have added....I
>>> might
>>> suggest that you not put a lot of emphasis or confidence on any one
>>> benchmark. Processors vary greatly internally, and so do benchmarks.
>>> It's
>>> not difficult to tune a benchmark to favor a specific architecture or
>>> configuration. This happens, and not always intentionally. Anytime we
>>> change
>>> something in a compiler or cpu config, etc., we run a half dozen or so
>>> benchmarks so that we get a better profiling of what the change did, and
>>> expose potential problems. No one benchmark can tell it all, and none of
>>> them are representative of what you will actually be doing with your
>>> computer.
>>>
>>>
>>>
>>>
>>> <miso@sushi.com> wrote in message
>>> news:e4934900-f01b-4282-a008-829053bb7823@59g2000hsb.googlegroups.com...
>>> > On Mar 28, 8:36 pm, Power Obsessed <screenedem...@comcast.net> wrote:
>>> >> I ended up figuring it out. If your interested it was the way the
>>> >> benchmark
>>> >> software was configured. I do want to thank all of you for being so
>>> >> helpful.
>>> >>
>>> >> Jim
>>>
>>>
>>>
>
 
Re: Quad core benching like a dual

Like I meantioned before the cpu load rarely exceeded 50%, so it is
unnessasary to O.C. my system. Like my screenname says though if the
potential is there I am going to use it to the point where I feel safe I
won't break it. Kind of like having a lamborgini and just idling around town.
--
Asus P5W DH , Q6600 Quad core 2.4ghz @3.12, thermalake V1 heatsink, 4gb
Crucial ballistix 800mhz , 16gb Mtron 6000 SSD, 150gb Raptor X, Nvidia
8800gt , 700 watt OCZ GameXStream power s , XP Pro X64


"Zootal" wrote:

> Desktop processors (pentiums, core, athlon, phenom, etc.) won't do it under
> load unless they are overheating. Lowering the voltage to unused functional
> units is an old technique, but they only do it when they think it's not
> needed. Specialty processors, OTOH, will do it even under load, depending on
> what it's currently doing. If we can scale the voltage down and still get
> the work done on time, then we do it. If I can distribute the load among two
> cores and lower the voltage and clock, I get the work done in the same
> amount of time, but use half the power. I don't need balls-to-the-wall
> processing 95% of the time.
>
> "Colin Barnhorst" <c.barnhorst@comcast.net> wrote in message
> news:44CD7C1B-C8D8-4E31-8D7C-3BC7412942B2@microsoft.com...
> > Modern processors throttle down to conserve energy and improve thermal
> > conditions.
> >
> > "Power Obsessed" <screenedemail@comcast.net> wrote in message
> > news:88425497-9428-4AC6-AC76-0A82D2FD275E@microsoft.com...
> >> To Zootal:
> >> The thing that really threw me was the HUGE difference in the results.
> >> (from 1600+ to 800+). I had used this software in the past "right out of
> >> the
> >> box" I guess you could say. No configuration at all.
> >> The setting I'm talking about is the number of processes to test the CPU
> >> with. I have no idea why that setting was different
> >> Thank,
> >> Jim
> >> --
> >> Asus P5W DH , Q6600 Quad core 2.4ghz @3.12, thermalake V1 heatsink, 4gb
> >> Crucial ballistix 800mhz , 16gb Mtron 6000 SSD, 150gb Raptor X, Nvidia
> >> 8800gt , 700 watt OCZ GameXStream power s , XP Pro X64
> >>
> >>
> >> "Zootal" wrote:
> >>
> >>> In addition to all of the varied comments that others have added....I
> >>> might
> >>> suggest that you not put a lot of emphasis or confidence on any one
> >>> benchmark. Processors vary greatly internally, and so do benchmarks.
> >>> It's
> >>> not difficult to tune a benchmark to favor a specific architecture or
> >>> configuration. This happens, and not always intentionally. Anytime we
> >>> change
> >>> something in a compiler or cpu config, etc., we run a half dozen or so
> >>> benchmarks so that we get a better profiling of what the change did, and
> >>> expose potential problems. No one benchmark can tell it all, and none of
> >>> them are representative of what you will actually be doing with your
> >>> computer.
> >>>
> >>>
> >>>
> >>>
> >>> <miso@sushi.com> wrote in message
> >>> news:e4934900-f01b-4282-a008-829053bb7823@59g2000hsb.googlegroups.com...
> >>> > On Mar 28, 8:36 pm, Power Obsessed <screenedem...@comcast.net> wrote:
> >>> >> I ended up figuring it out. If your interested it was the way the
> >>> >> benchmark
> >>> >> software was configured. I do want to thank all of you for being so
> >>> >> helpful.
> >>> >>
> >>> >> Jim
> >>>
> >>>
> >>>
> >
>
>
>
 
Back
Top