Firing Squid Wastes their Ink

FrgMstr

Just Plain Mean
Staff member
2FA
Joined
May 18, 1997
Messages
58,704
Firing Squid Wastes their Ink - Our reply to an attack on real world testing. Some will have you believe that a canned benchmark or gaming timedemo is a crystal ball into your gaming future.

Please discuss.
 
The nice thing about actual benchmarking is that you're able to see how the different aspects of hardware and visual quality scale with each other. I can see where the bottlenecks are from the benchmarks, and make a judgement based on the hardware I own. The real world evaluations are only worth while if the reader happens to own exactly the same hardware setup (other than the one piece being reviewed) as the reviewer.
 
I certainly agree. But the exact same can be said of canned benchmarks. But what makes canned marks even worse is that even after taking into consideration differences in hardware it still in no way reflects I would get even if I DID have the same hardware as the reviewer...
 
hughJ said:
The real world evaluations are only worth while if the reader happens to own exactly the same hardware (other than the one piece being reviewed) as the reviewer.

"Real world" benchmarks would require the same hardware as well, would they not? I don't understand what you just said. :confused:
 
I have to agree with FiringSquad on this one.

Here’s a question for you, after seeing CPU test benchmarks that show identical performance between three processors because the video card is the bottleneck, what have we learned about the CPU’s performance? How do we know how it compares to other processors in its class? Moreover, those tests indicate nothing about what we can expect from it if it is paired with a better video card.


I don't even completely agree that the "real-world" testing doesn't give a good evaluation, but it does definitely leave out some more of the "hard facts" that let people compare for themselves instead of trusting someone else to do it for them. However, with that said, it must ALSO be said that whatever method a site wants to use is fine... variety is good :D! Let us see both and decide based off of EXTRA information that way.
 
I have no problems with [H]ard|OCP's Core 2 review. It was different than other reviews and
different is fine by me. [H]ard|OCP is always my first place to go when I am looking for a
good review on a piece of hardware, but I always get a 2nd, 3rd and sometimes 8th opinion.

What I don't like about the [H]ard|OCP review was the accusations of "lying" and "BS". Also
the title of this thread is a tad unprofessional. I believe that is called "name calling" even
though it is rather clever.

Nothing can shake my love for the [H] though. I just received my [H]ard|OCP t-shirt and soon
I am going to sign up for the lifetime subscription to General Mayhem.
 
I would have to vote in favor of "artificial benchmarks. It is a rule of science, involve the fewest variables possible. Realworld benchmarks are great for perspective, but I cannot see the artifical ones goin "out of style" anytime soon.
 
Kyle, I hate to play blacksheep as I have been a reader of Hardocp a long time I recall even seeing you on Techtv once if I remember correctly. Let me get it strait that I do not think your data is false. It is just incomplete. You may hate the low res way and timedemo ways but there an easy standard to go by if everyone is using them we dont have to worry about each site using a different one. You have been around hardware a long time, as long as me more than likely as Im one of the older readers here. You know that when your trying to isolate a specific component you remove all other bottlenecks.

The entire Gaming performance review was bottlenecked which is nothing new really. It just proves if you have an athlon FX 62 or a new conroe without SLI your hitting the max it can go. Throwing in some low res marks and adding a variety of games and settings would of made the review very sound. But as a technology lover Im not partcial to amd or intel I just buy the best available to me.However, I felt parts of the review smacked of biased opinion instead of a un partial look into conroes performance..

Thanks
 
Just tell everyone you were testing how games performed on the Core2. That's all you have to say. If you just say that Kyle, people will realize what you were doing. Then point them to your other articles for the rest of the testing. People don't understand that you don't game at 640x480.

However next time I would throw in a few of those low resolution benchies just to make ppl happy.

Remember Kyle, just keep telling everyone they are gaming benchmarks and then point them to the other articles.
 
i think you guys are missing the point. the other Core 2 performance evaluations on [H] showed more gains. but when it came to gaming, brent and kyle just called it like it is--there is not much of a gain to be had from going to an fx-62 to the conroe chip. the point of the article was to provide a gaming experience evaluation, not to provide a cpu speed benchmark with all confounding variables removed.

what i took away from the articles is that with the current generation of games, there are not much gains to be had by upgrading to the intel chip. when it comes to encoding and number crunching performance, it's a different story.

i appreciate the [H] giving me a way to see what kind of experience i can expect with new hardware. canned benchmarks may tell me what cpu / video card / ram is faster in very limited and unusual conditions, but when it comes to playing games on my computer, i care more about the experience and that's what the [H] evaluation philosophy is about.

there are plenty of other sites for canned benchmarks. you guys are arguing against the wrong thesis.
 
I think you are both ([H] and F/S) wasting your ink and time. I see both methods (real-world and canned, 3DMark etc) as beneficial for hardware evaluation, and as stated here, others seek multiple opinions first. It would be foolish when purchasing such high-cost items to do otherwise.

My opinion on this mess? Seriously, when you make a blanket accusation that is stereotypical in nature, you may stir up a hornets nest and have to take a few lumps in the process. But I'd just like to see it all move forward.

Kudos to [H] for innovation and promoting a new way of thinking, but let's not just trash the old methods just yet ;)
 
I believe the problem with synthetic benchmarks is this:

I read website #A, they show a low resolution "gaming performance benchmark" lead of 50 fps using traditional benchmarks. I buy the CPU, based on that 50 fps lead....

...only to realize that (as a real world evaluation showed) there is no massive lead for real world gamers, playing games like normal people play them (not at 640x480).

Obviously my opinion doesn't count in this discussion but there it is anyhow ;)
 
Canned benchmarks tell you one thing only: How products compare to each other in running that benchmark. If your goal is to get better timedemos or more 3dmarks or a better SpecINT rating, then by all means, use a canned benchmark.

Some people have trouble interpreting real world benchmarks. A few posters in here have mentioned that real world benchmarks only apply if you have the exact same hardware as the reviewer. I submit this is even more the case with canned benchmarks. Do you really know what a canned benchmark is telling you? Does a high score mean you'll be able to play games better?

Obviously with [H]OCP's Core 2 review, that is not the case. Canned benchmarks would lead you to belive that you should trash your A64 and run out and get a Core 2.

Real world evaluations require significantly more investment of a reviewer's time and mental resources. "Jakub's Rant" spends most of his time talking about how a real world benchmark is more difficult to execute, and how it's more difficult to interpret the results.

Jakub also points out, wisely, that bottlenecks present a barrier to understanding the true performance of a component. While this is true, I would say to Jakub that yes, your 500x faster processor is quite nice, but if it spends all day long waiting on the other components in my system, what kind of performance benefit do I get? Zero.

He also asks "When these next-gen video cards come out, how will you be able to choose the appropriate processor to match them, if all your CPU benchmarks were run with older video cards that are maxed out?" Thats easy - you wait for the next gen card to come out, and you wait for [H]OCP's review of that new card on the new CPU. In this case, his logic means that he is targeting one audience only: The extreme early adopter.

If someone rushes out and buys a Core 2 right now because of what the next generation of video cards might be able to get them to do, he may get locked out by other technology. The next gen of video cards could have electrical or signaling requirements that the current Core 2 mobo/cpu can't handle. Early adopters understand this, or they should, and they pay the price.

Jakub obviously has trouble grasping the importance of different types of metrics for value.

Kyle on the other hand, seemed a little too threatened by Jakub's article, responding in similar tone rather than maintaining his usual professionalism. The reference to a "well trained monkey," while true, was not professional. The facts speak for themselves.

People like canned benchmarks for the same reason they like high megapixel cameras with awful quality. Its easier to understand a number ("16k 3dmarks is better than 12k" == "8 megapixels is better than 6") than a subjective analysis. Dan recently updated a really good article (thanks digg!) that talks about megapixels (http://www.dansdata.com/gz059.htm) and I feel like this is really the same type of discussion.

-MC
 
When was the last time any reputable review site Included ONLY low res numbers though ? None of the ones I read. My two main sites Hothardware.com and Anandtech.com
 
hifi said:
I think you are both ([H] and F/S) wasting your ink and time. I see both methods (real-world and canned, 3DMark etc) as beneficial for hardware evaluation, and as stated here, others seek multiple opinions first. It would be foolish when purchasing such high-cost items to do otherwise.

My opinion on this mess? Seriously, when you make a blanket accusation that is stereotypical in nature, you may stir up a hornets nest and have to take a few lumps in the process. But I'd just like to see it all move forward.

Kudos to [H] for innovation and promoting a new way of thinking, but let's not just trash the old methods just yet ;)

I second what he said. Educated enthusiasts will be able to make their own decisions on performance levels based off different benchmarks / real world evaluations. If everyone used the same systems and same testing methods, there would be no reason to search incessantly for performance reviews. And that would take away all the fun.
 
FlyinBrian said:
It is just incomplete.

Thanks

FlyinBrian, I agree with you in one sense: Sometimes canned benchmarks give us an easier reference to understand relative performance.

[H]'ers: how about hiring a "Well trained monkey" to spit out canned benchmarks to supplement reviews? Many people on here have mentioned that they hit 3 or 4 other review sites to get those canned benchmarks. If your systems are already setup for the real-world analysis, a canned benchmark won't take significantly more time to perform and shouldn't detract from the subjective analysis much at all.

The real value for a canned benchmark is very little, but if your readers want it, they will spend more time interacting with your site.

-MC
 
Kyle and Crew,


Really, at the end of the day, it's all about balance I think.

I think you should find the sweet spot of hardware testing, some real world and some simple canned time benches.

The people that come to Hardocp are smart people. Most if not all are going to understand what you put in front of them.

Personally, I am more concerned with name calling than they method in which hardware is tested here.

Be the bigger website and ignore other peoples drama. You do not need to point out or prove anything. You guy are doing a fantastic job here.

Keep in mind, no one wants pages and pages of hardcore details. In fact, some of the graphs you guys use ugly to look at.
 
Zinn said:
i think you guys are missing the point. the other Core 2 performance evaluations on [H] showed more gains. but when it came to gaming, brent and kyle just called it like it is--there is not much of a gain to be had from going to an fx-62 to the conroe chip. the point of the article was to provide a gaming experience evaluation, not to provide a cpu speed benchmark with all confounding variables removed.

QFFT. The point is not that the Core 2 ISN'T better than what's out there -- just that it's not really better for current gaming. When you consider the other testing done that shows the Core 2 does indeed put the beat down on AMD in applications other than gaming, there's no real justification in bitching about a reality-based gaming eval. H is the number one resource for gaming hardware reviews for me, because I actually care about what new hardware does in REALITY. To me, canned benchmarks are much like the minimum requirements listed on a game: you aren't getting what you're being shown at that level. Why do I care what a cpu will push for fps when you're talking about the lowest common denominator in regards to IQ? Nothing. Nothing at all. H has it right when it comes to gaming evals.
 
Well name calling another web site gives you the truthiness upperhand I guess...

When an NDA expires, I refer to all the major web sites to try and form a rough conclusion. It's not perfect, but I have a darn good idea after 30 mins of clicking. What I do find tacky is the tactic voicing condescending remarks towards other sites, which helps to produce more google bucks (click advertising) than it does to solve the problem.

Anand keeps it clean and whether or not you approve of his methods, his site is the Prince of hardware reviews. HardOCP and Tomshardware do far too much bickering. While Firingsquad may not have the hardware street cred of the afore mentioned sites, I go there for the thorough game reviews and fairly frequent hardware round ups on current games.

Let your data speak for themselves and don't hate in order to prove your point. Just state them, rant if you like and disagree, but don't become a bottom dweller and berate other sites with derogatory comments.

Some comments why I don't like the hardocp format of benching.

1) A wiggly line means crap to me. Give me the average. I've yet to see a game where two opposing hardware setups produced wiggly lines that crossed, and resulted in an invalid average. How many wiggly line charts have you guys made where the upper line wasn't consistently 2/3's above the next highest hardware setup? The average still comes out to the higher card being first. "Ooh wiggly line at the 34 sec mark, that must mean you made a turn at the fountain and met the final boss, which obviously is drawn using SM2.0b, and that caused a drop." No one will deduce that level of granualarity from your wiggly lines.

2) I have no idea why you continue to make tables with numbers showing the min/max/ave.

eg. Min 30 Max 50 Ave 45.

http://enthusiast.hardocp.com/article.html?art=MTEwOCw3LCxoZW50aHVzaWFzdA==

is a great example.

Can anyone tell me why you aren't using a bargraph? Produce three bargraphs and put the title at the bottom describing the results.

...

Anyways hardocp/firingsquad do good reviews and sometimes iffy...Critique, but lose the attitude...cause we know it's only to get you more google bucks.
 
My goodness. I didn't know being able to have 1000+ fps at 640x480 with no eyecandy could really mean this damn much to anybody. Or a 100K 3DMark score. (Yes, those are extremely exaggerated numbers. That would be the sarcasm.)

Kyle, and Steve, there's only one thing i have to say about this. My grandfather told it to me a long time ago:

An empty can makes the loudest noise.

The best reply you can give to Firing Squad is simple - nothing. Let them say what they want. Your evaluations, your work, your reputations, and most importantly, your success, speaks for itself in your defense. As long as you have that going for you, Firing Squad can toot their horn all day, it's not going to mean a lick of difference to you, and shouldn't.

Proof in example: There are people out there who tout that their AthlonXP's can stomp anything out there now. I wish I was kidding on that, but...unfortunately I'm not. People will believe what they want to believe, even if they have all evidence pointing to the contrary. You can't convince them of otherwise, so don't waste the effort. Save your energy for the people who are willing to listen.
 
SamuraiInBlack said:
My goodness. I didn't know being able to have 1000+ fps at 640x480 with no eyecandy could really mean this damn much to anybody. Or a 100K 3DMark score. (Yes, those are extremely exaggerated numbers. That would be the sarcasm.)

Why the hell do people keep bringing up the 640 x 480 point??

Where in Firingsquad's review of Conroe do they even test at 640? I see a setting for 800 x 600 and 1600 x 1200.

http://www.firingsquad.com/hardware/intel_core_2_performance/page12.asp

I've seen this numerous times throughout various threads.

Are you just quoting Kyle's core2duo? So damn annoying.

Firingsquad's results that I just linked are IN LINE with Kyle's results that a single GPU is a bottlenecker and so many CPUs will run about the same speed.

Afterwards, they go into a crossfire analysis:

http://www.firingsquad.com/hardware/intel_core_2_performance/page15.asp

which shows the benefits of the processor.

So canned benchmark or not, each site has produced evaluation results that make sense.
 
I agree with Kyle's philosophy. It doesn't matter to me that at 800x600 you can get 200+ FPS with the latest CPU because I don't play at 800x600. Just think how many people would be upset if the high milage car they purchased was discovered to have conducted it's MPG testing completely stipped down with things like the doors and seats removed. Of course it's going to get better milage if it weighs less, but that isn't how you are going to drive it now is it? Test it the way it's used, we accept that philosophy for many other products, why should PC components be any different?

Given that, what I would like to see in future [H] reviews is more data points. For example, the Core 2 Duo reviews only compared two Core 2 Duo CPUs against a single FX-62. That's great if you are trying to show how the top-end of Intel compares to the top-end of AMD, and naturally when comparing top-end CPUs, the margins of performance difference are more likely to be smaller. What I'd like to see is comparisons to other, more commonly owned CPUs, so that even if my own CPU isn't represented, I could pick one that I knew to be "close" in performance for my own personal comparison of what the upgrade might mean for me. From the Articles [H]ardOCP wrote I know that the Core 2 Duo is clearly faster than what I own, because I know that an FX-62 is faster than what I own, but I don't have a clear idea of how much faster, which really matters when trying to justify to the wife the expensive purchase you want to make. More data points that show how the Core 2 Duo compares to older CPUs would have made these articles perfect.
 
Triple B said:
QFFT. The point is not that the Core 2 ISN'T better than what's out there -- just that it's not really better for current gaming. When you consider the other testing done that shows the Core 2 does indeed put the beat down on AMD in applications other than gaming, there's no real justification in bitching about a reality-based gaming eval. H is the number one resource for gaming hardware reviews for me, because I actually care about what new hardware does in REALITY. To me, canned benchmarks are much like the minimum requirements listed on a game: you aren't getting what you're being shown at that level. Why do I care what a cpu will push for fps when you're talking about the lowest common denominator in regards to IQ? Nothing. Nothing at all. H has it right when it comes to gaming evals.


It is better for current gaming no doubt about it unless you have an athlon FX 62 and then the gpu is already bottlenecked if you do. But for people who have lower end to mid range processors some may read the review, and many have in this forum start spouting off that intel lied about conroes performance in gaming, which they did not; then proceed to say i knew they were lieing ETC ETC . They just used sli to showcase their products power over their competition.
 
There are two sites I hit up daily for hardware news myself. HardOCP first, and Anandtech second.

Those that say variety is good I have to agree with. When I'm plumping down money that, on my college budget, ends up being a sacrifice in many other areas I like to be fully educated on the matter. I like to see both "real world" benchmarks, such as HardOCPs, and "canned" benchmarks, purely as a matter of showing raw performance which *MAY* hint at performance in future games. (Not necessarily though, as I would imagine that as complexity increases an architecture that once did great could get bottlenecked in an unforeseen way)

For whatever its worth I prefer HardOCP's way. A lot of times there are "apples to apples" pages within reviews which goes to show the old-school way to supplement the more modern approach of appraising quality over quantity. A consumer would simply be remiss to not go elsewhere, at least three or four sites, reading technical overviews, etc, before making a purchase. [H] is my first stop, and when I've looked at others, I come back for a final look.

However.. I think Steve and Kyle, along with Brent of course, are great guys going out of their way to try to innovate in the American style by bringing a quality product to the market. If one doesn't like it in those terms, then they're trying, in the open style of Democracy, to bring quality journalism.. but being an economics major, and since it's not a non-profit site, the product analogy works best. I think the crew here should let their fine work stand on its own and take the high road when little boys whine and cry and bitch and sling names around like children and not pay them the respect of response. I believe an article such as the CPU Scaling one stands on its own and visitors to the site can clearly, after thinking about it, see the value of the different reviews posted here. On the same token, calling out the other sites as "liars" might've been.. provocative. Needlessly so, perhaps. [H] would earn more respect, I think, but outperforming peers instead of lowering itself to their level. Not [H]'s fault they're doing reviews that, apparently, are getting stale! It's called innovation, and if other sites had it they'd do well too. I can understand getting pissed off though when these jokers post a hot-headed rant a mile long which basically amounts to slander, I'd probably talk back as well, but then again I don't try to run a site viewed by thousands.

Anyway, bottom line: Keep up the good work. People that complain need to understand they can get their other benchmarks here and getting everything from one source is ignorant and irresponsible as a consumer anyway. And taking the high road would just look better.. probably not feel as good, but it'd look better. ;)

Edit: As an aside question, Kyle, or anybody else that knows: Does [H]ardOCP purchase their products on the market like the rest of us, or take samples? Or a mixture of both, with some disclosure? Either way I think is fine, but +3 for credibility if, in the tradition of Consumers Union and the likes, they pay for it themselves instead of feeling any obligation to make a review turn out one way or the other.
 
just glanced at a few posts here..inbetween shows.

but FSquad included multiple approaches to their review (which they admit is a cpu review :).

this site tested only at one extreme res that 3% of gamers use as already proved by linkage to steam survey. We will say 21% of people here use that res..just to avoid argument.

But will add..these same high end gamers...are prolly also the same few percent that use sli to extend their capabilities.

thus, this review was limited and intended to only show one thing ...as kyle stated in that other thread:

"still got the messae across that I was looking to get out. Conroe does not give you huge gaming benefits as has been hyped by Intel and their cronies for months..."

But yes, we need to see more than one res. and more than the high end chips...so everyone can see how they stand with their current systems.

of course, if you had just upgraded...why would you want to upgrade again so soon.
But for the vast majority of the other gamers...

:)
 
Jeez... I have so much input on this subject but limited patience with my fingers on the keyboard.... so I will sum-up my opinion on "real-world" testing the best I can....

If you need to get an idea of how a new proc technology will compare to a competitors or your own hardware then the only thing you should look at is the "real-world" tests (HardOCP Style)... who cares if the ut2004 benchmark(with bots) runs better on core duo with the res at 800x600 ?? who the fuck is gonna play the game like that ??? I wanna know if my hard earned cash is gonna be well spent by upgrading to a newer tech... if the real world results indicate a 2fps increase ... then NO i'm not gonna invest

Now with that said I feel that if you want to compare raw proccesor speed then "canned" benchmarks can work very well... if thats what you feel others will judge your E-penis size by the Raw speed of YOUR shiney new intel core duo.. then go with a review from the firing squad....

To me comparing the way these two different sites compile thier reviews is pointless... IF YOU WANT TO KNOW REAL WORLD RESULTS... READ HARDOCP (or others that do testing the same way)

IF YOU WANT TO WHIP OUT THE RULERS THEN GO SOMEWHERE ELSE... I'm done
 
RedStarSQD said:
this site tested only at one extreme res that 3% of gamers use as already proved by linkage to steam survey. We will say 21% of people here use that res..just to avoid argument.

On the one hand, Steam includes grandma's 1999 computer that the idiot 10yr old grandkids installed Counter-Strike Source on.. but on the other..

Sure I lower the res or turn down the quality in multiplayer, in a more taxing game.. but do readers of a site like this REALLY still use 1024x768?!? I haven't used any less than 1600x1200 on my desktop since the month the Dell 2000FP came out, later bought my current 2001FP after the original got killed. If my card can at all handle 16x12 I use it, if not, I'll step down to the res [H] reviewed but no lower, ever.

Are there really that many people here reading about overclocking X2 3800+'s to 2.8ghz and FX60's and the latest 7900 GTX MADMAN XXX69 THUNDERDEATH 512mb DDR9-II using 1024 when playing SP for fun?
 
To expand: The way we test video cards is to use them in the same way that a gamer uses them, thus providing real-world feedback on performance and image quality and how video cards compare in those regards. Moving this method of testing to CPUs seems logical for today's CPUs and games.

I think we made the right decision to do this now because it really shows how CPUs compare when using it in a real-world everyday gaming situation just like you'll use it at home. Playing through games and figuring out what settings are playable on a system is what everyone does before they play a game. We do that exact same thing which tells you which video card or which CPU is better and what their relative gaming experience is like. Our method realistically represents gamers playing games.

The scientific method no longer applies to evaluating video cards and CPUs for gaming. It may apply for other types of evaluations, but not for determining real-world gaming experiences. When you want to figure out the gaming experience a product delivers there is no other way to do that IMO, and that's exactly what we are trying to represent, the real-world gaming experience.

That's my peace.
 
A: firingsquad creates an artical to FLAME another site
B: Hard OCP responds to FLAME with their own FLAME
C: Hard OCP creates a thread in its forums to FLAME firingsquad

did i use the word enough? thats exactly what this is

why would you respond? have the bloody deceincy to ignore it. your acting like the people in your forums you try so hard to weed out. i didnt think a website... that does reviews or hardware... could sink so low.

and you guys think that because you throw in some differances in AA between cards that your tests are so much better then theirs?

testing Core 2 under in low resolutions simply exaggerates the results. its like putting something under a microscope. sure your never gonna play at 800X600 (OH BTW is it possible that someone buys core2 on a budget and can only get 800X600? oh no dont consider that, thats not apparently "real world" enough for you, maybe this is mainly a workstation but would be nice to play wow? perhalps s/he has to use integrated with his/her core 2 proc?) i thought we were all educated enough to realize this. guess i was wrong.

this is repugnant.
 
Tzzird said:
Why the hell do people keep bringing up the 640 x 480 point??

Where in Firingsquad's review of Conroe do they even test at 640? I see a setting for 800 x 600 and 1600 x 1200.

http://www.firingsquad.com/hardware/intel_core_2_performance/page12.asp

I've seen this numerous times throughout various threads.

Are you just quoting Kyle's core2duo? So damn annoying.

I fail to see what the eyecandy settings are other than HDR for Oblivion. I see that they're specified for Quake, and just about all the other graphs, and yet not Oblivion. Why? I don't own the game to know anything about the eyecandy levels or how to set them in-game, but if I was going to play a game as immersive as Oblivion is touted to be, I'd certainly like to have my eyecandy, especially if I'm throwing $1000-$3000 or more into a machine to get it. An X1900 XTX Crossfire setup? That's hardly a mom-and-pop setup there just on video cards alone, nevermind the CPU of choice. If it is spelled out I'd get better eyecandy down at a hellhole strip joint, well then, I guess i know where my paychecks are going to go to for the next 2 years or until there's reasonable price drops.

And yes, I checked the system setup list to see if it was listed there as well. I feel that finding out what eyecandy level settings are capable of in having an effect on performance should also be a determining factor, not just on CPUs, but hardware in general that can affect visual quality. Maybe eyecandy is GPU specific, but personally? I want proof of that. I'm not saying that it isn't, but I am also saying I won't believe that it is, until I see proof beyond doubt of that.

The 640x480 point isn't a jab at that particular resolution, but at the fact that anybody who is really going to invest that kind of cash into a system, is NOT going to be playing at low resolutions, PERIOD. If I had the kind of cash for a Core 2 and all the parts associated with something that expensive to invest in, I wouldn't be gaming at anything less than 1280x1024 myself, and that would probably be a laughable thing for some around here who think 1600x1200 is the new bare minimum.

To me, the lack of specifying settings, canned benches or not, is just bad writing, and makes for a bad article. If I wanted bad writing, I'd IM my best friend's ex girlfriend and try to interpret it.
 
He had to respond to something, he cant respond to the idiots on digg. Everyone is in an uproar about this and squid was the easiest to respond to.
 
The original article was a flame in itself. I'm going to blame who started it: H.
 
CaptRingold said:
Edit: As an aside question, Kyle, or anybody else that knows: Does [H]ardOCP purchase their products on the market like the rest of us, or take samples? Or a mixture of both, with some disclosure? Either way I think is fine, but +3 for credibility if, in the tradition of Consumers Union and the likes, they pay for it themselves instead of feeling any obligation to make a review turn out one way or the other.

Recently some articles have been sponsored by BFG. Which was another question mark brought up as to why they didnt try CrossFire setup after failing with the SLI setup.

All lot of questions were asked in the other thread that were never fully answered or avoided.

One that was brought up is why no Simulators or RTS were tested. Driving or Flight. They would strain the CPUs more than a FPS or action adventure game.
For a "REAL WORLD GAMING BENCHMARK" it was a narrow view.

Maybe [H] should survey their readers first instead of just saying Single GPU systems = 1600x1200. 1280x1024 systems will likely see a 10-20fps differance enough to take a stuttery game to being playable if you already have the GPU and are looking for a CPU upgrade. Which is more "Real World" than 1600x1200 with fluctuations of 20-30fps making GAME TEARING much more likely.
 
CaptRingold said:
On the one hand, Steam includes grandma's 1999 computer that the idiot 10yr old grandkids installed Counter-Strike Source on.. but on the other..

Sure I lower the res or turn down the quality in multiplayer, in a more taxing game.. but do readers of a site like this REALLY still use 1024x768?!? I haven't used any less than 1600x1200 on my desktop since the month the Dell 2000FP came out, later bought my current 2001FP after the original got killed. If my card can at all handle 16x12 I use it, if not, I'll step down to the res [H] reviewed but no lower, ever.

Are there really that many people here reading about overclocking X2 3800+'s to 2.8ghz and FX60's and the latest 7900 GTX MADMAN XXX69 THUNDERDEATH 512mb DDR9-II using 1024 when playing SP for fun?


I really doubt people on here really play at 1024 nowadays... I haven't used anything lower than 1280 even for AGES, let alone 1024. With purchasing a $1000 CPU, you would hope you weren't dense enough to put it in with a $200 graphics card and assume you're going to have a fast rig. I think bottlenecks are pretty obvious to anyone, and variables inherently mess up testing and invalidate it. It's like gravity... if you have nothing supporting something in thin air, it is going to fall, there is NO question of it. It's fine to say you want to do real-world testing, but to still call it scientific when it's really more opinion is just a bit much in my opinion.

Let's make a comparison:

One site: Car A offers a smooth driving experience going 0-60. It handles it well, and on the road surfaces I usually drive on while racing, provides an optimal speed when you press the accelerator down at 90%. However, Car B, the reigning champion, provides the same experience. I think that there is no point in selling your old Car B and switching to Car A.

Another: Car A offers a great performance, going 0-60 in 4.2s. Handling is great, and on standard pavement achieves the performance levels indicated. On the surfaces I race on in drags, however, it doesn't do so well. With both the Car A and the reigning champion, Car B, you will get great performance, no question. However, Car A has more raw power under proper testing conditions, and will have room to grow as your tires change and are upgraded to provide better traction.


See the difference? YES, both are VALID AND PROVABLE, but one is an opinion, the other is a scientific fact. I just don't want one presented as though it's the other.
 
MrWizard6600 said:
A: firingsquad creates an artical to FLAME another site
B: Hard OCP responds to FLAME with their own FLAME
C: Hard OCP creates a thread in its forums to FLAME firingsquad

did i use the word enough? thats exactly what this is. a fucking pissing match between morons.

Actually, it goes A. HardOCP makes an article with an opening flaming all other sites' testing methods, B. Firingsquad flames HardOCP back in their response, C. HardOCP responds with their own argument, and then D. HardOCP makes a thread in its forums to help support their argument :).
 
MrWizard6600 said:
A: firingsquad creates an artical to FLAME another site
B: Hard OCP responds to FLAME with their own FLAME
C: Hard OCP creates a thread in its forums to FLAME firingsquad

d.

you forgot the real A: [H] makes an article and blasts all the other sites needlessly.

B: Fsquad responds ....
C: this thread
D: No doubt other sites ("cronies of intel") will also weigh in at some point
 
Back
Top