Derek Wilson of Anandtech...

nelg said:
Panzeramd's conspiracy theories aside, the only thing that could be construed as bias was this remark.

At the time that statement had a lot more teeth as we had seen some pretty remarkable performance increases pulled together through driver evolution. NVIDIA did have some fairly sizable jumps IIRC.

On the same note, I am fairly certain we have used the same basic line in regards to ATI. Their driver program has evolved greatly over the last couple of years and we have seen tremendous performance increases over certain products lifetime.

On both sides, this is why we retest so much equipment every time a new driver is released. A new driver can paint a new picture of a product for sure, be it red or green in color.
 
JoshMST said:
Good grief.

Let me let you in on a couple of "little secrets" about online journalists...

We are human, we make mistakes. Unfortunately, because we lay everything we have on the line, those mistakes end up being posted on forums with wonderful comments like "******" "biased" "ignorant" "stupid" "liar" along with all these theories how we are being paid off one way or the other. While there may be some journalists that do receive some form of payment or another, that is far from being the status quo.

As journalists, we have contacts with these companies. These contacts are human too, and they have their own likes and dislikes. Because a PR person might personally like a journo, they often are friendlier to them, try harder to get them stuff, lead them to potential advertisers, etc. This is not to sway that journo, but rather it is a favor born out of friendship. I have several PR people that I am good friends with, that we discuss things far away from graphics/cpus/motherboards/etc. But business is business, and if one of these companies really screws up, I owe it to my readers to tell it how it is. Hard, AT, and others have all been equally hard on whatever IHV at one time or another, but it seems that many are saying these guys are green without really thinking why. I think the biggest reason that favor has switched over to NV is that NV hasn't stumbled since the FX days, while ATI has had quite a few issues in the past two years straight. I for one am ecstatic to see ATI come out with the X1900, and they have executed flawlessly on it. This is a big change from the past two years. I think we will start seeing that favor swing back to ATI as they have shipping product, while NV can't produce enough GTX 512's to satisfy demand, and they are still several months away from having an answer to the R580.

Journos may also not be getting the entire story, so their point of view may in fact be skewed. This skewed outlook is not some malicious artifact of having bias or being paid off. A good example is when I wrote that horrid State of 3D several years back. The main reason I didn't have much good information about ATI's products is that ATI was not willing to talk to me at that time. I learned a whole lot by producing that steaming pile, and while I am embarrased by the content of that article, I think in the end it was a very positive experience because it did create a lot of discussion, I learned a lot about both NV's and ATI's architecture, and ATI finally started to respond to me. So often when people call out bias on a journo, most often that person just has been exposed to really one side of the story. PR people have it tough as well, because there are thousands of would-be startups that put up a website and want hardware/information/contacts.

Whenever I hear bias, I always think back to the saying, "When you point a finger at someone, you are pointing three right back at you" (work with me, visualize pointing at someone with one finger... hopefully the index). Derek made a mistake, big deal. That doesn't mean that he sits in his office and is hand fed information directly from NV headquarters, and doesn't think for himself. Very few journos know anything about process tech other than just the numbers that are thrown out. In fact, how many users at large had any clue that he even made that mistake? Its a tough crowd here.

As for GTX 512 and 7300... NV continues to ship those big cards, but obviously not in good numbers. The issue with NV is getting chips that will run at that speed (not a PCB issue, or memory availability). Qualifying G70 parts to run at 550 MHz at 100C is not easy I am guessing. 7300's are now available from eVGA at their site, and from Newegg and others. More availability will be soon. Honestly, I could give a rats if the 7300 was hard launched, because who is really going to buy it? The hardcore enthusiast who can't wait to get his hands on this technology? I think not. This is going to be mainly a business class card and an OEM favorite. If it hits retail two weeks after launch, big deal. Now, if NVIDIA releases their next high end card to no availability, then we can all scream "paper launch" and they deserve to be castigated for it. That will be a high demand part for the very vocal enthusiast crowd, and if NV can't deliver, then there will be hell to pay. But for the 7300?

So just remember, us journos on the whole are not bad people. We don't take bribes, but we are also human in that if we are treated well by a certain company (namely they talk to us, which honestly is not that common) we may also cut some folks a break. Overall though, we are pleased to keep companies to task, as that is our job.

From the perspective of the public, there are critical problems with the online computer hardware reporting.

Example:
Because a PR person might personally like a journo, they often are friendlier to them, try harder to get them stuff, lead them to potential advertisers, etc. This is not to sway that journo, but rather it is a favor born out of friendship. I have several PR people that I am good friends with,...
It's their job, Josh. Reading the above makes me a bit sad, because I've heard PR people talk. They have a personality that's amenable to their job, but - getting you to feel like they honestly like you is their job.

Another angle they use is making you feel as if you're "part of the industry". Which is partly true. The problem is that "the industry" in their vocabulary are the purveyors of kit. Not the actual end users.
The ultimate "interest of the industry" is flogging the maximum amount of goods to the public at the highest possible profit margin. It behooves the rest of us always to remember that.

Consumers have very little to help them, the (online) journalists and articles being pretty much the only source of "independent" information available. Thus, they are a powerful means to manipulate the buying public. That's why the benchmark cheating is so ugly, because it tricks people who might otherwise be honest to stand up and speak your lie for you, and it also takes away the only real source of information that the public can use to draw their own conclusions from.

A fundamental problem is the truthful part of the reporters being part of the industry. Hardware IHVs, printed magazines, web-sites... have a strong common interest - all benefit from retaining public mind-share. Something has to be percieved to be continually going on. Hardware IHVs play the numbers game with their product (under how many names has the design powering the Radeon 9600 been sold? 10+?), they make minor adjustments of clocks, or these days, turn on/off functional units et cetera. Hardware sites test these, make statements, and so on. It is all part of making a buzz, because for the most part true advancement comes from silicon lithography, and those changes are on a much too long time scale to keep consumer interest up. And lately the benefits haven't been as big as they used to, either.

So - anything that comes along that is even vaguely interesting is going to get covered. Witness the relevance attached to SLI for instance. At best, an inefficient and expensive way to gain another notch of resolution. There are NO statistics available anywhere that show how many consumers actually assemble such systems. But that doesn't stop websites to always use such set-ups as points of reference in their comparison charts. Which was obviously nVidias idea. And the sites played right into their hands. Which forced ATI, which pressured VIA..... Where was the consumer interest in all of that? Aren't the consequences and costs involved that get pushed to all consumers actually a negative? But there is precious little reporting that brings up that line, because a fundamental weakness of hardware reporting is co-dependence. Everybody wants stuff to happen, and so all plays are overall going to get a more positive response than they deserve. Does the public have any reason at all to be enthusiastic about the new chips using 32-bit FP rather than 24-bit FP? It is presented to us (why should we care at all?) as a positive, but why not mention the costs in transistors and thus die size, cost and power draw (=clock speed) for what seems to amount to no visual benefit at all?

The examples are less important than the principle - those who sell kit and those who report on those products to the public have significant common interests.

Being chummy with the ones you're supposed to report on just drives the nails in all the deeper.
 
I was actually interested in SLI way before Nvidia even launched it. As soon as PCI-E express spec became public, I noted it will allow more than one card. I don't see offering higher end options to consumers as a negative. More choice is better. The same goes for Platinum/Extreme editions in reviews.

I don't see anything wrong with reviewing every little feature, SKU, etc that comes out. These are options for the consumer, and how else would anyone know how a new clock-bumped SKU compares unless someone reviews it. SLI is a consumer option. Lots of white collar 25-40 types have disposable income sitting around to spend on this, just like lots of people blow money on their cars, and home consumer electronics equipment, for marginal gains in performance.

You seem to be arguing that journalists should write maybe 1-2 stories a year when new chips are taped out, I think that's a pretty dumb idea. I also think that regardless of PR people shipping kits around, people would be reviewing every little SKU that came out anyway. The only difference is, the reviews would be written by end users who actually went out and bought the HW for personal use, vs "professional" online journalists.
 
DemoCoder said:
The only difference is, the reviews would be written by end users who actually went out and bought the HW for personal use, vs "professional" online journalists.

Except when end users do that, and you* don't agree with them - even though you* don't have any experience with the product - you* blast them as being biased.....

*collective you, not meant personally ;)

I actually agree with you on the review part, and even on the idea of multiple GPUs. But, so far, I've been pretty dissappointed in SLi...and I DO own an SLi system.
 
Entropy, obviously there are a few things that I need to address.

I have been doing this reporting/reviewing thing for about 7 years now, so I have more than a little bit of experience here.

First off, yes, it is the PR's jobs to get the word out on their product. They need to do it in a professional and friendly manner. Now, while I deal with dozens of PR folks, I honestly can say that there are about 5 that I am actual friends with. This is not me wishing and hoping that "they will be my friend" but these are people who have gone well beyond their job descriptions in helping me, giving me information not normally passed on, as well as giving their honest opinion of the situation (and often times this opinion is not exactly positive towards the product or company in question). You are basically insunuating that I am some poor, friendless loser who takes in PR spin hook, line, and sinker. I have to disagree with you. PR and reporting are two jobs that interact with each other, and while it is almost always cordial (except when huge controversies happen and there is a lot of finger pointing) reporters typically know which PR people they can trust, and which ones are just "doing their job". Again, of the 50 or so PR folks I tend to work with every year, I honestly like and trust about 5 of them.

Now, there is give and take between the manufacturers and press. In fact, if anything, the press is more reliant on the manufacturers than vice versa. Manufacturers get news of their product out through the press, at the same time the press get stories and product reviews from the manufacturers. If there were no press, the manufacturers would still sell product. The other way around doesn't exactly work...

I think a good recent example of a positive change due to press pressure was that of the X1600 series. They were initially going to retail the high end product at $249, but there was pretty much universal outcry that it was far too expensive for what it offered. Now we see these top end X1600 products retailing for under $160 US most of the time, and for most of us that is a good price/performance ratio.

"The examples are less important than the principle - those who sell kit and those who report on those products to the public have significant common interests.

Being chummy with the ones you're supposed to report on just drives the nails in all the deeper."

I really take offense at this comment. While we do have common interests, and in many ways we are two dogs eating at the same bowl, but to throw a blanket statement saying that we are basically in collusion with each other to hoodwink readers and users is pretty insulting. I think that the majority of journalists out there covering the industry will call a pig a pig. I mean, look at the near universal outcry from journalists about the Intel Pentium4 and how it is so hot, pulls so much power, but can't get past the AMD products in performance. Intel has billions they spend in advertising, and they send out a lot of "kit" to reviewers. Do you often hear us say, "Yeah, the P4 isn't so bad, and I would recommend it over the Athlon 64"?

Here is the deal... reporters are far more responsive to our readers then we are with manufacturers. Our readers are our bread and butter, and it doesn't matter how much material we get from manufacturers, if we don't have readers we soon become inconsequential. Once that happens, everyone abandons us. No readers = no advertisers, and no traffic = no manufacturers willing to send "kit". So, while we may have a cozy relationship with manufacturers (source of products and information) without readers we would be nothing. To keep readers we have to develop trust with them, and to do that we can't afford to act as shills for these manufacturers. I know of a couple of places where the reporters/editors did in fact "sell themselves out" and those sites are now marginalized and have little traffic.

Entropy, you seem like a very pessimistic person, I am curious... where are you from? Don't get me wrong, I think a healthy dose of skepticism is a good thing, and it keeps us from being fooled all the time. But you present your argument like there is some kind of conspiracy between reporters and manufacturers, and that every PR person is a malicious entity that's only reason for existing is to promote the good of their product over every other consideration (mind, their job is to promote their product, but I hardly think they are evil for doing that). I am curious about your background.
 
JoshMST said:
I really take offense at this comment. While we do have common interests, and in many ways we are two dogs eating at the same bowl, but to throw a blanket statement saying that we are basically in collusion with each other to hoodwink readers and users is pretty insulting.

Yes, that would be pretty insulting...however, that's not what he said at all. He's not saying there is some overt conspiracy between journalists like yourself and manufacturers. And quite frankly, your last post would only heighten his actual concerns. Please read his post again.
 
I understand his concern, and I see where he is coming from, but I guess my viewpoint is one that really has to be experienced through a reporter's eyes.

First off, I disagree with his argument that advancements in lithography is the only really interesting or impressive change from generation to generation. GeForce 3 was a 150 nm part, and the Radeon 9800 XT was also 150 nm. I would say there are some significant and exciting changes between those two products using the same process node. Granted the 150 nm process was two years old by the time the 9800 series was fabbed on it, but the amount of changes were not that great (just a better mix and more optimal yields). I think process is certainly helpful and fascinating, but I think that silicon design is probably more important than just the process jumping we see going on.

Journalists also need things to write about, just as you and I need to breathe. If we put out two or three articles a year, then we would find a quick end to our careers. Also, whenever a consumer heads to Newegg, Best Buy, or wherever, they are greeted with shelves full of products that may or may not have any real differentiation. There is a need for the information that we publish (whether it be about the top of the line video cards, or if someone wants to know if buying a X1300 would be better than getting a 7300 GS).

I also think your aversion to SLI (which is pretty common around here) is actually somewhat interesting. I do not see SLI as a vital component, and in fact I much prefer just having one video card in my machine (noise, power, heat). However, SLI is an option, and it should be looked at as merely an option. It honestly is something that is unpalletable for a large portion of users, but as long as there are people that want such a setup, S3, ATI, and NVIDIA will provide multi-GPU solutions. This is simply an example of our semi-free market economy. If there is a demand for it, these companies will provide the products for supply. Nevermind that a good portion of games will never see any benefit from it, or that there could be issues with the rendering, it is simply a market option. I mean, who really needs a Ford GT or Ferrari F460 to go back and forth from work? I certainly don't, and I don't find either of those the least bit economical sounding, but there are obviously people who buy them. (Dammit, you made me do a car analogy!)

So, I sorta get what you are saying, but at the same time I am confused because what other system is there that could work? People need 3rd party information about products so they spend their money wisely, these 3rd parties need to interact with the manufacturers to get information from the source as well as review products. Any interaction leads to interpersonal communication between the reporters and the manufacturers. The only way I could see getting around this is if a website charged each "member" about $50 a year to subscribe to the publication, and the publication goes out and buys all the necessary hardware to review and report on, all the while staying away from the manufacturers. The problem with that is how many people would subscribe to such a publication? Sure, lots of people subscribe to magazines, but that is a medium that is starting to die. We are all very addicted to the idea that we can visit websites and get information from free.

Other suggestions and thoughts would be welcome.

Digi: I did not mean to give insult, and I was not trying to imply that he was part of some group or mindset that was directly opposed to mine. I think the issue we might be running into is that we are using a very imperfect medium (forum posts) to try to express our point of view, and sometimes it is hard to get the little things explained well. It could also be that we both have such a different point of view that we cannot fully express what is important and what is trivial for each of us. Again, I am sorry if that post did appear condescending or insulting, it was not my intention. Rather, I was just curious as to his background so I could maybe really get a finger on what his main thrust of his argument is.
 
Last edited by a moderator:
JoshMST said:
I mean, who really needs a Ford GT or Ferrari F460 to go back and forth from work? I certainly don't, and I don't find either of those the least bit economical sounding, but there are obviously people who buy them. (Dammit, you made me do a car analogy!)
/me quickly throws a big sheet over our 350Z!

Uhm, yeah...well I still think dual video cards are a fad.
bleh2.gif
 
JoshMST said:
I also think your aversion to SLI (which is pretty common around here) is actually somewhat interesting. I do not see SLI as a vital component, and in fact I much prefer just having one video card in my machine (noise, power, heat). However, SLI is an option, and it should be looked at as merely an option. It honestly is something that is unpalletable for a large portion of users, but as long as there are people that want such a setup, S3, ATI, and NVIDIA will provide multi-GPU solutions.

It would be nice if you and others actually expressed this in your reviews. Fact is, most sites really do make SLi seem like something that is a must have. You seldom see anything posted in reviews about it's drawbacks - besides cost. I really appreciate your post...I just wish more "reviewers" would admit how they really feel.
 
martrox said:
It would be nice if you and others actually expressed this in your reviews. Fact is, most sites really do make SLi seem like something that is a must have. You seldom see anything posted in reviews about it's drawbacks - besides cost. I really appreciate your post...I just wish more "reviewers" would admit how they really feel.

I dunno, it seems to me most reviews I see say don't bother if you don't have a monitor that can go above 1600x1200. . . which describes a pretty hefty percentage of the user base. Tho both IHV's tried to address that a bit with SuperAA and SLIAA.
 
Xmas said:
How do you know they don't write how they really feel?

I don't know, and I never in any way intended to imply that I do. But, if you look at all of the major sites and many of the minor ones, whenever SLi is brought up it seems none of the disadvantages are even mentioned, except the cost. I really do appreciate not only what JoshMST said, but the way he said it.

geo said:
I dunno, it seems to me most reviews I see say don't bother if you don't have a monitor that can go above 1600x1200. . . which describes a pretty hefty percentage of the user base. Tho both IHV's tried to address that a bit with SuperAA and SLIAA.

I agree......and will add that most all also say that you need serious CPU power also. ;)
 
JoshMST said:
First off, I disagree with his argument that advancements in lithography is the only really interesting or impressive change from generation to generation. GeForce 3 was a 150 nm part, and the Radeon 9800 XT was also 150 nm. I would say there are some significant and exciting changes between those two products using the same process node. Granted the 150 nm process was two years old by the time the 9800 series was fabbed on it, but the amount of changes were not that great (just a better mix and more optimal yields). I think process is certainly helpful and fascinating, but I think that silicon design is probably more important than just the process jumping we see going on.
Josh, while it´s certainly not "the only" real interesting thing about GPUs, it´s one of those doors that keep closed, whenever possible. Since silicon design goes hand in hand with lithography, there is no such thing as "more important".

So, actually, it´s a combination of both design and process targets, since you don´t design your GPUs regardless of those, which are depending on a lot of factors, therefore you can´t make such "simple" sounding (150nm vs. 150nm) comparisons like that, because they are not. As already stated, process improvements and knowledge gained from a node is another well kept secret, since those things cost all IHVs plenty of time and money. TSMC made available at least 2 targets fairly late in the game, at a point where 150nm was very mature, which allowed ATi to squeeze everything out of their original R300-design, so, when talking about the comparison you just made, 150nm is not the same 150nm when Geforce 3 was designed, actually it´s quite the opposite.

Speaking of "advancements", they can can make fairly huge differences, when talking about the whole picture. So your take on it is a little bit "off base", i´m afraid.

You can do a search on that if you want, it has been discussed many times.
 
Last edited by a moderator:
EEEK! Remind me to not speak in such general terms when I post on this board!

That entire portion was based around "I disagree with his arguments that advancements in lithography (node changes) is the only real interesting thing around GPU's.

Yes, when looking at the tools creating the GeForce 3 and then comparing them to what was used at ATI with the Radeon 9800 series, there are many years of differences and the engineering experience from both sides helped a great deal. No argument there. I also look at what was done with the first 130 nm products and the final products on the 110 nm process. Again, huge leaps. I guess I again should have been more explicit in my post, mainly that the node changes are not the most exciting things about GPU's and their technology... but I already thought my previous post was long enough.

As far as I am concerned, and please judge my respect to the engineers behind such products, but that any of this even works (and works as well as it does) is just PFM to me (Pure F*****g Magic). What they do with silicon based transistors is just amazing.
 
Back
Top