Open Question & Recommendations for Tomshardware Editors

Ok, I realize some of the recommendations have been around the forum already but I will take the time to again list them with better detail. Also sorry for posting this in the wrong section. I do hope the TH editors are looking at the Forums.

=============

Before I do the recommendations I would like to ask a few questions most of these haven't been asked before as fas as I know.

1. I would like to know how you measure temps for CPU/GPU/Motherboard. If you are using IR thermometers please list model. If using software please list the software. I am asking this due to the work done by CompuTronix and the author of RealTemp. Their work showed that the temperature read by CoreTemp and other software are inaccurate for some if not all 45nm CPUs.

2. Also when the OCing is done on factory sample do you double check the work with retail versions of the part?

3. For OCing CPUs is the OCing done through the BIOS only or do you use software like EasyTune (Gigabyte OCing software) or the likes of that? Also are the BIOS flashed to the latest version for the motherboard(s)?

================

Now for the recommendations (1= highest recommendation):

1. Tomshardware NEEDS to do a high quality PSU round out. By this I mean taking apart the PSU and looking at the parts, loading the PSU with a high drain, testing the voltage drops on each line,etc. Kind of smiler to the ones done by jonnyguru. Including a mix of brands from low to high end would also help.You should include PSUs that some people buy based on cost like the Rosewill and Raidmax. Also include the higher end PSUs from Corsair, Antec, PC Power & Cooling, OCZ, etc. A very good PSU round up would (imo), include at least 25 PSUs. Putting this in a interactive chart (based on voltage,etc) smiler to the CPU chart will also help greatly.

2. Tomshardware needs to re-do the CPU Cooler Chart (possibly for 2009). The 2008 HSF chart was not very good. You should NOT base a HSF on how hard it's to install the cooler. That is an OPINION not a fact therefor it is not very good to base a chart on an opinion. You can comment on the installation but it should not influence the rating of the cooler. The cooler should be rated on noise(as in measured in dB) and cooling performance. Frostytech dose very good cooler reviews and doing some thing smiler will greatly help. If creating a chart it should be based o the cost, cooling performence, and the noise of the HSF.

3. Tomshardware should do a LCD display round up (~20+ LCds from low to high end). There has never been one and doing so will help a lot of people.






 

amdfangirl

Expert
Ambassador
They need to make everything a scientific test... Only ONE independant variable if possible... not charts were motherboards differ... sure I agree benchmarking is a bore but we need fairness?
 

randomizer

Champion
Moderator

Some do, some don't. Most probably don't and just read article comments. I'm not sure who is meant to read this section to be honest.

+1 Good idea!

I don't quite understand what you want here. Are you asking if they check to see if they can overclock the same on retail samples?
As to which is better, most say the BIOS, but I say that is debatable to an extent. I can't get passed the loading screen at >3.5GHz without BSODs even if I run at 1.75V (max for my board), yet I can run prime95 for several minutes (probably longer but that isn't necessary since a Windows boot isn't that intensive). There could be another failing component at play though during bootup.

*Getting tired of quoting*
Nobody likes the current chart format, that should be worked on (and I think it is) first before the charts are actually updated/added to.


If only they'd done it before I bought my LCD :(


Rebenchmarking all those systems for all those benchmarks would be extremely time consuming, not to mention some of those CPUs probably won't even work on the current test platform, especially when they make a Core i7 ( :pfff: ) rig.
 

amdfangirl

Expert
Ambassador
^ Nah, keep the old stuff...

Auto benchmarking FTW!



Lolz, just got my new 21.6" and 40" =P and a HD 2600 PRO =(

No silly, I meant keeping as many controlled variables as possible ie. same motherboard if possible...

Its easy to quote when you just grab the entire chunk... then nobody realises what you are saying!