Benchmark tests for OC...

Do you have a question? Post it now! No Registration Necessary.  Now with pictures!

Threaded View

New here.  Just wondering if anyone knew of any good downloadable
software to test my benchmarks.

Posted using the interface, at author's request
Articles individually checked for conformance to usenet standards
Topic URL:
Visit Topic URL to contact author (reg. req'd).  Report abuse:

Re: Benchmark tests for OC...


Re: Benchmark tests for OC...

'AngryPigeon' posted:
 "New here.  Just wondering if anyone knew of any good downloadable software
to test my benchmarks."

What exactly do you wish the benchmarks to do?

If you want to see the change as it affects the operation of the CPU and
memory, and as a guide to increasing overclock performance, then 'SiSoft
Sandra 2004' ( ) will show the effect of
overclocking on different aspects of memory and CPU subsystems (L2 cache,
memory bandwidth, SSE instructions, floating point) with the tests 'CPU
Arithmetic Benchmark', CPU Multi-Media Benchmark', 'Memory Bandwidth
Benchmark', 'Memory and Cache Benchmark.')

If you want to see the improve of the overall performance of your system,
the best test of that is how the performance 'feels' to you in the
applications you actually use.

If you want to compare overall system performance, be aware that there are
many subsystems that aren't affected by overclocking (hard drives, removable
media drives, display adapters [which can sometimes be overclocked
separately]), so a 30% overclock of the CPU and memory may effect a much
smaller increase in overall system performance.  For this kind of comparison
the benchmarks used by review sites like 'Tom's Hardware' ( ) and 'Anand Tech' ( )are useful, but many of these benchmarks are
hundreds of megabytes in size, and mainly useful for bragging rights.

By the way, I recommended 'SiSoft Sandra 2004' even though there is a newer
version "SiSoft Sandra 2005' available.  That's because, while there is a
free, but limited version of each available, the 2005 free version is not
quite as useful.  The only reasons for using the 2005 version are
(1)  you can't find the 2004 version on the web
(2)  you want to compare your overclock performance to that of other

'SiSoft Sandra 2004' is simple to use, gives rapid results, and is a small
download.  No matter what information you wish to gather, 'SiSoft Sandra
2004' is a good starting point.  (note:  'Sisoft Sandra' has had weaknesses
in the past, but it has improved over the years;  the low-level benchmarks
are good, it's just that some of the other types of information that 'SiSoft
Sandra' gathers might have some inaccuracies.

Just like overclocking, start small and work up, results vary according to
your patience.

Phil Weldon

Quoted text here. Click to load it

Re: Benchmark tests for OC...

First benchmark test is that of stability :-)

o BIOS screen with temp reading is useful
---- in most cases there is no power mgt re CPU
o Memory test in particular is very important
---- prove timings now, not with data later
o HD testing should not be ignored
---- re cooling using A/V scan or Defrag
o CPU/GPU testing
---- stability trading voltage / frequency
---- temperatures at peak - ideally hot environment

Secondly, measurement of gains :-)
o Matching gains to usage, eg, MPEG to Games

Typically it is worth stepping back from the most
severe settings where stability is of more concern.
Dorothy Bradbury for NMB & Panaflo (MMM) fans, free shipping

Re: Benchmark tests for OC...

'Dorothy Bradbury' wrote, in part:
"First benchmark test is that of stability :-)"

A good suggestion, to start with stability tests.
Your post has suggested to me that BASELINE results of all tests should be
gatherred before any overclocking is attempted.

Phil Weldon

Quoted text here. Click to load it

Site Timeline