redgarl said:
For anyone that want to know what matter...
https://static.techspot.com/articles-info/1870/bench/Cost.png
With the actual scoring, this review is a joke. Navi is disrupting pricing and almost match a 1080 TI for 400$... however at tomshardware it sux.
While I was reading this article I was thinking "please please please let redgarl post in the forums!!" My wish was granted and Redgarl definitely provided.
Redgarl describes this review as a "joke". He claims that Navi is "disrupting pricing" and "almost match a 1080 TI". He provides links to the TechSpot 5700 review for "anyone that want to know what matter..." This made me genuinely really curious. Is today the day that redgarl proves Tomshardware is run by a bunch of hacks? I had to know for myself.
First of all I looked at the test setups for each review. Techspot uses a 9900K Intel CPU and 32GB DDR4-3200. Tomshardware's build is different. They use a 8086K Intel CPU and 64GB of DDR4 2400. Tomshardware documents the drivers used as 431.16 Nvidia Driver for the 2060Super and 2070Super but the 430.86 for every other Nvidia card. For AMD cards, they use the 19.7.1 driver with the RX 5700 cards and the 19.6.3 driver for everything else. By comparison, Techspot reports that they used the "latest drivers available at the time". Its not clear how that compares to Tomshardware. The test machine on each respective site is different and that may cause differences in benchmarks. I predict (jk. I've already seen the results) the Tech Spot benchmarks may be a little higher judging by the faster memory and the better CPU.
I then compared Techspots benchmarks to those featured on Tomshardware. First I looked at the Assassin's Creed Odyssey results. At
1440p, Techspot reports the Radeon RX 5700XT hit an average of 70fps. Meanwhile, Tomshardware was showing... oh dear, Odyssey wasn't featured on Tom's. Let's just move on.
The Techspot's Destiny 2 1440p results show that.. uh oh, it's happening again. Techspot didn't feature Destiny 2 benchmarks. This is a problem because you can't compare apples-to-apples when the same benchmarks aren't used.
Assassin's Creed Odyssey, Destiny 2, DiRT Rally 2.0, Far Cry 5, Far Cry New Dawn, Final Fantasy XV, GTA V, Resident Evil 2, Strange Brigade, Tom Clancy's Ghost Recon, Tom Clancy's Rainbow Six Siege, World War Z, The Witcher 3, Wolfenstein II: The New Colossus are either benchmarked by Tom's or TechSpot but not both. The only overlapping games are Battlefield V, Forza Horizon 4, Metro Exodus, Shadow of the Tomb Raider, and Tom Clancy's The Division 2.
Let's look at the Shadow of the Tomb Raider results. Techspot and Tom's ran the game at 1440p, highest quality.... oh no. Tomshardware used SMAAT2x in it's Shadow of the Tomb Raider benchmark. Techspot didn't report using that setting. Well that's not apples-to-apples either. A similar situation happens with Battlefield V where Tomshardware uses DX12 and Techspot uses DX11.
For Tom Clancy's The Division 2, I double checked to make sure the software is the same between both Techspot and Tom's platforms. DX12, 1440p, Ultra quality. Check. Now here is something that appears to be genuinely inconsistent between reviews. Techspot's averages for the GeForce 2080, 2070 Super, 2060 Super, Radeon RX 5700, and 5700XT are higher than the Tomshardware numbers. Well there you have it.
Redgarl has proven that "With the actual scoring, this review is a joke." Tomshardware losses, redgarl wins. Except that the test systems between Tomshardware and TechSpot are different. Isn't this a roller coaster? TechSpot has a better CPU and faster ram. I looked carefully at the trends between both reviews and lo and behold they show the same thing. Both sites list the video card pecking order as: 2080, 2070 Super, Radeon VII, 1080 Ti, 5700XT, 2070, 2060 Super, Vega 64, 5700, 1080, 2060, and so on and so forth. It's very plausible that the differences in specific numbers can be chalked up to differences in platforms or margin of error.
The charts that redgarl cherry picked show the results between Toms and Techspot are totally different. If you only considered those charts then you might think Tomshardware has an incompetent staff the way redgarl wants you to believe. However, when you look at the whole picture, its obvious that you can't directly compare the benchmark results between both sites. A total of 19 games were reviewed by both sites. Only five overlap. I threw out Shadow of the Tomb Raider because we can't be sure both sites used the same settings. That leaves four. When we directly compare results, the numbers vary probably due to differences in test hardware, but the trends are consistent. In conclusion: there isn't enough evidence to prove that the Tomshardware review did anything wrong.
To redgarl, this is just for fun my dude. No hard feelings.