This has to be the most controversial launch of a GPU since the 5800 Ultra. HardwareAnalysis has certainly created a firestorm by releasing a set of benchmarks that doesn?t show the type of performance one would expect from the new ATI architecture.
The type of reaction to this event is well expected. A hardware site, that wasn?t invited to the ATI R520 launch and isn?t under the NDA, releases R520 benchmarks that weren?t done in house and shows the R520 in a poor light. We all know that more than a few NVIDIA loyalists would use this as fanboy fodder, while at the same time some ATI loyalists would slander the hardware site and the writer of the article and call its benchmarks ?fake?. In other words, the response was rather typical.
The more neutral fans of NVIDIA/ATI took these benchmarks with a grain of salt. Benchmarks from one source, no matter how scientific it was done or how much we trust the site, can never be weighed alone. Benchmarks from a number of different trusted sites should be used to make a valid judgment. We all know that mistakes are made and that issues are often overlooked when our trusted hardware site are rushing to produce performance articles to be released as soon as the NDA expired. It can be assumed that more errors can be introduced when product comparisons are made by two different testers (with each tester testing only one of the competing products but not both) and using two different test beds (this is the case with HA benchmark review). Our opinion of the R520 performance should not be swayed by these benchmarks, simply because we cannot form an honest unbiased opinion until more data is in hand.
While the graphics community?s reaction to the pre-released benchmarks would well expected, ATI?s was not. The problem with this fiasco isn?t the so-called R520 benchmarks, but ATI?s response to them. Their statements and their handling of the situation have compounded the doubts created by the benchmarks. While I expected a statement from ATI, I thought I would be reading something like this:
?Dear So and So Readers,
While we understand that HA has not signed an NDA and is no way bound by its restriction. Furthermore, the production of pre-release benchmarks of R520 would be newsworthy event for any non-NDA bound hardware site. However, we do not agree with the questionable benchmarks produced by HA or the dubious conditions in which they were attained. We call into questions the reliability of the benchmarks due to a number of issues.
1.The benchmarks numbers are not indicative of the performance of the X1800XT. The performance testing done in house produces results that contradict the HA?s benchmark numbers.
2. No board partners have access to the R520XT or its specs. We are still in the process of finalizing the specification of the R520XT to ensure that we reach the maximum performance potential of our flagship card. Therefore, with no R520XT or its specification, it is highly improbable for a board partner to reliably produce the performance characteristic of the R520XT.
3. While we cannot prove or disprove Mr. Sander?s account of how these benchmarks were attained. We feel that our business partners would not risk their relationship with ATI just to provide ?performance material? to a largely unknown hardware site.
While we would like to release performance numbers that contradict HA?s analysis of the R520, it would be to the dismay as well as detriment of the trusted hardware sites that are bound by the NDA. Where one of their prominent functions is to provide unbiased and objective performance reviews for many of ATI products. Those sites have agreements with ATI to release no material related to the R520 specifications and performance characteristic in exchange for access to the R520. We expect them to honor the NDA, so we feel that we should honor the Non Disclosure Agreement as well.
Thank you,
So and So
ATI?
However, what we got was statements and other material that would make one believe that the PR team of ATI needs a class in Damage Control 101. While they couldn?t disprove or prove the statements and data given by Mr. Sanders, we ended up hearing statements about ?over-clocked 850XT, we think?. We were given the gist of a bunch of email fodder that cast doubts Mr. Sanders? character. However, those gists don?t even show circumstantial evidence that might make one conclude that Mr. Sanders may have concocted that data. It is just mostly used as a tool of character assassination. This tactic gives credibility to Mr. Sander?s assertion all along, that ATI doesn?t play kindly to those who write unflattering articles of ATI products. ATI response seems very desperate, vindictive and unprofessional. Instead of limiting the damage caused by the unflattering pre-released benchmarks, ATI fueled the flame.
This controversial article would have died smoldering death with a very simple and professional response saying it doesn?t believe that these benchmarks can be used to determine the performance potential of the R520XT. Nevertheless, ATI response has insured that HA?s articles will a topic of conversations in a lot of community forums until after the expiration of the NDA. Also, ATI has left themselves in difficult spot. If by some chance, these benchmarks are true, ATI will probably lose a lot of credibility in the eyes of the consumers.
Disclaimer: This view is not to disparage the R520 or any other ATI product or levy an opinion of whether the 7800 will be better than the R520 or vice versa. This is again is just something I'm doing to fill the time until we see if the R520 is the glorious return of the 9700, a 5800 in disguise or something in between.
The type of reaction to this event is well expected. A hardware site, that wasn?t invited to the ATI R520 launch and isn?t under the NDA, releases R520 benchmarks that weren?t done in house and shows the R520 in a poor light. We all know that more than a few NVIDIA loyalists would use this as fanboy fodder, while at the same time some ATI loyalists would slander the hardware site and the writer of the article and call its benchmarks ?fake?. In other words, the response was rather typical.
The more neutral fans of NVIDIA/ATI took these benchmarks with a grain of salt. Benchmarks from one source, no matter how scientific it was done or how much we trust the site, can never be weighed alone. Benchmarks from a number of different trusted sites should be used to make a valid judgment. We all know that mistakes are made and that issues are often overlooked when our trusted hardware site are rushing to produce performance articles to be released as soon as the NDA expired. It can be assumed that more errors can be introduced when product comparisons are made by two different testers (with each tester testing only one of the competing products but not both) and using two different test beds (this is the case with HA benchmark review). Our opinion of the R520 performance should not be swayed by these benchmarks, simply because we cannot form an honest unbiased opinion until more data is in hand.
While the graphics community?s reaction to the pre-released benchmarks would well expected, ATI?s was not. The problem with this fiasco isn?t the so-called R520 benchmarks, but ATI?s response to them. Their statements and their handling of the situation have compounded the doubts created by the benchmarks. While I expected a statement from ATI, I thought I would be reading something like this:
?Dear So and So Readers,
While we understand that HA has not signed an NDA and is no way bound by its restriction. Furthermore, the production of pre-release benchmarks of R520 would be newsworthy event for any non-NDA bound hardware site. However, we do not agree with the questionable benchmarks produced by HA or the dubious conditions in which they were attained. We call into questions the reliability of the benchmarks due to a number of issues.
1.The benchmarks numbers are not indicative of the performance of the X1800XT. The performance testing done in house produces results that contradict the HA?s benchmark numbers.
2. No board partners have access to the R520XT or its specs. We are still in the process of finalizing the specification of the R520XT to ensure that we reach the maximum performance potential of our flagship card. Therefore, with no R520XT or its specification, it is highly improbable for a board partner to reliably produce the performance characteristic of the R520XT.
3. While we cannot prove or disprove Mr. Sander?s account of how these benchmarks were attained. We feel that our business partners would not risk their relationship with ATI just to provide ?performance material? to a largely unknown hardware site.
While we would like to release performance numbers that contradict HA?s analysis of the R520, it would be to the dismay as well as detriment of the trusted hardware sites that are bound by the NDA. Where one of their prominent functions is to provide unbiased and objective performance reviews for many of ATI products. Those sites have agreements with ATI to release no material related to the R520 specifications and performance characteristic in exchange for access to the R520. We expect them to honor the NDA, so we feel that we should honor the Non Disclosure Agreement as well.
Thank you,
So and So
ATI?
However, what we got was statements and other material that would make one believe that the PR team of ATI needs a class in Damage Control 101. While they couldn?t disprove or prove the statements and data given by Mr. Sanders, we ended up hearing statements about ?over-clocked 850XT, we think?. We were given the gist of a bunch of email fodder that cast doubts Mr. Sanders? character. However, those gists don?t even show circumstantial evidence that might make one conclude that Mr. Sanders may have concocted that data. It is just mostly used as a tool of character assassination. This tactic gives credibility to Mr. Sander?s assertion all along, that ATI doesn?t play kindly to those who write unflattering articles of ATI products. ATI response seems very desperate, vindictive and unprofessional. Instead of limiting the damage caused by the unflattering pre-released benchmarks, ATI fueled the flame.
This controversial article would have died smoldering death with a very simple and professional response saying it doesn?t believe that these benchmarks can be used to determine the performance potential of the R520XT. Nevertheless, ATI response has insured that HA?s articles will a topic of conversations in a lot of community forums until after the expiration of the NDA. Also, ATI has left themselves in difficult spot. If by some chance, these benchmarks are true, ATI will probably lose a lot of credibility in the eyes of the consumers.
Disclaimer: This view is not to disparage the R520 or any other ATI product or levy an opinion of whether the 7800 will be better than the R520 or vice versa. This is again is just something I'm doing to fill the time until we see if the R520 is the glorious return of the 9700, a 5800 in disguise or something in between.
Facebook
Twitter