Do you think reviewers REALLY STRESS systems enough?

[TantrumusMaximus]

I love the information here at Anandtech. But I sure would love it if reviewers as a whole across the board would stress boards more than they do. I don't mean length of time but putting a REAL enthusiast load on the board and see how it reacts.

Something like:
4 DIMMS
2 Optical Drives
2 IDE drives
2 SATA drives
2 PCI cards

Populate the system and Overclock it. Let us know how they react then. Not the typical 1 card, 1 hard drive, 2 dimm setup. Not only would we get feedback on how the Motherboards reacted but how PS's reacted. The x reference PS may not mesh well with x mobo etc etc. I know it's impossible to give us every situation possible but it would be nice to see a little more in the test beds.

 

[Wicked2010]

For what purpose?

I don't see the point of doing this.

 

[HeaterCore]

I think it's better to go barebones and show the baseline overclockability of the board/chip. If reviewers start stuffing test systems with random gear, they might unfairly hold back a part because of some random incompatibility that most end-users -- with different gear -- would never see. With the barebones approach, at least we know where a part can go.

-HC-

 

[CaBoOse999]

The results would be irevelant because not everyone has the same hardware.

 

[TantrumusMaximus]

I guess this didn't come across right.

I'm not talking from a performance perspective, yes by all means get some stats how the board performs "speedwise" against others. It just seems that we get gleeming reviews on a part and then run out and buy the board and find OH look if I have this many parts it doesn't like it. It is NOT uncommon for people to have mult HD + optical drives. It's not that uncommon to populate your DIMM slots.

I would like to know before I buy that the board in question responds well to these conditions.

I know there is not a perfect test, but I think EACH and EVERY feature of a MOBO should be analyzed for a good review. If a board has 4 DIMM slots... fill em, all the controllers should be tested with a drive on each of them simultaneously.

It just seems that people only seem to care about one thing... speed and overclocking and seem to overlook the other stuff.

 

[VirtualLarry]

I think that you are talking about stability testing under max (power, bandwidth, etc.) load. That's a good idea, but I have a feeling that they may not have access to the hardware long enough to perform those sorts of tests, along with the ones that they do now.

 

[DAPUNISHER]

Too many variables@play with reviewed hardware more often than not, egineering samples, beta bioses, that kind of stuff. IMO that limits how useful the information is to what I can expect from the retail board I will get from e-tailers anyhow. Let them do their thing, use it as a basic guideline for what board does what, but don't develope high expectations based on what you have read. They pimp this stuff for a living, ya know? My point is, YMMV regardless of what/how they test.

 

[PhoenixOrion]

Originally posted by: DAPUNISHER
Let them do their thing, use it as a basic guideline for what board does what

then look at the hundreds of posts in cpu and motherboard forums how this part is the suxxors and how this part is the roxxors = informed buying decision from numerous retai-market samples with various configurations.

 

[DAPUNISHER]

Originally posted by: PhoenixOrion

Originally posted by: DAPUNISHER
Let them do their thing, use it as a basic guideline for what board does what

then look at the hundreds of posts in cpu and motherboard forums how this part is the suxxors and how this part is the roxxors = informed buying decision from numerous retai-market samples with various configurations.

:thumbsup: Even the forum posts need to be carefully scrutinized, as there are many n00bs and boobs having a PICNIC < *trademarked M4H I look for posts by members with excellent testing methodology and the necessary experience to properly evaluate the hardware before taking the pros&cons seriously. But the principle you promote is the most sound means of making an informed purchase, no doubt in my mind.

 

[superkdogg]

If you add all the other hardware, it becomes a PSU stress test and shouldn't change results anyway. I think that most people correctly assume that optical drives and PCI cards perform virtually the same regardless of motherboard. Most reviews at least touch on performance of onboard controllers-especially if they are not up to snuff.

If you're looking for system-wide performance numbers, use the ORB at some benchmark makers' website. Otherwise, the motherboard can be expected to perform the same with more attached hardware as it does with little attached.

Kind of along these same lines, I like the benchmarks because they show how a 2.4c NW, and a 2400+ TB performed. This demonstrates the real performance differences that an upgrade brings rather than simply showing that most new stuff is really cool and fast and performs pretty similar to its competition.

***WARNING- That statement was a gross, broad generalization. Please do not flame about architecture, etc. Yes, there are differences and I know. But generally, new stuff is much better and a 3.8 Prescott and FX-55 both destroy a TBred about equally.

 

[TantrumusMaximus]

Originally posted by: VirtualLarry
I think that you are talking about stability testing under max (power, bandwidth, etc.) load. That's a good idea, but I have a feeling that they may not have access to the hardware long enough to perform those sorts of tests, along with the ones that they do now.

Even so...even the PSU reviews I read across the net don't seem to really have the resources to really stress the PSUs. Placing a PSU in a 2 HD system and saying yeah it's great! The digimeter readings all looked great and then just barking all the features off and giving it an "<Insert Website Stamp> Choice" award isn't really a good review. To really stress a PSU for a review I think it should be placed in a system that would load it to what it is actually rated for and perhaps beyond and see how it performed... under good and bad thermal conditions.

Motherboards.... I agree a true MB review to please everyone is kinda tough. That's why these forums are great...but like said above you have to take the real data vesus the noob data and decide for yourself what is real. I still believe that reviewers can stretch them a little further or even as said above give me some benches on not only the latest greatest CPU but tell me the numbers on another lower end proc too.

The reviews can be better.

 

[Jeff7181]

The memory test would be nice, but I think we all know that a lot of A64 boards have trouble when you fill all the DIMM slots... can't run the fastest timings.