JAMES GORBOLD / HARDWARE ACCELERATED
James Gorbold has been building, tweaking and overclocking PCs ever since the 1980s. He now helps Scan Computers to develop new systems.
OPINION
U
nfortunately, becausetwoofthewordsinmyjob
title are ‘TechnicalMarketing’,I sometimesgetdragged
into projects aspectsofwhichdrivemearoundthe
bend. One such recent projectrequiredustoproposea new
range of systems and submitbenchmarkresultsatspecific
settings. The purpose oftheprojectitselfisnotimportant;
what is was the franklyridiculoussettingsatwhichwehad
to run the benchmarks.
Part of me was relievedtoseethatthebenchmarkswereall
real games, and not synthetictestssuchas3DMark.However,
two of the games werefrom2016,withonlyonefrom 2018
and none from this year,sononeofthem
was really going to pushtheboundaries
of modern hardware, letalonegetgamers
excited about recent advancessuchasDXR.
What upset me the most,however,was
how once again I was beingaskedtoruna
load of benchmarks at 1,920x 1,080.This
would have been excusableifwewere
speccing up entry-level PCs, but even the lowest spec model
had a GeForce RTX 2070 Super, a graphics card that can deliver
smooth frames at 2,560 x 1,440. Now this might seem a trivial
matter to get annoyed enough about to write a column, but
such ridiculous benchmarketing creates two serious issues.
The first problem that benchmarking high-end hardware
at a low resolution creates results that are near useless as a
guide to help customers make their buying decision. Instead of
providing useful information such as what resolution monitor
the system is optimal for, benchmarking high-end PCs at such
a low resolution merely creates results that demonstrate that
all such specs are ridiculously overpowered for 1,920 x 1,080.
Forexample,comparingtwosystems,the first around £2,000
incVATandthesecondaround£3,000 inc VAT, were 150fps
and180fps.Inotherwords,bothsystems were capable of
equallyridiculouslyhighframeratesat 1,920 x 1,080.
The second issue testing suchhigh-end hardware
configurationsatlowresolutioncreates is that it encourages
themyththatcertainCPUsarecritical for gaming. While
it’sfairtosaythere’sa quantifiableperformance difference
betweenthelatestIntelandAMDCPUs at 1,920 x 1,080, once
youcrankuptheresolutiontoa level the system will likely
beusedatintherealworld,theperformance difference is
negligible.I getthat it’s important to test
productsfairlyand on a level playing field,
butfindit dismaying that even to this day,
somereviewerscontinue this practice and
propagatethismyth.
I’mnot,tobeclear, accusing everyone
ofbeingdisingenuous or engaging in
benchmarketing,as I suspect the root cause
for most of these low-resolution benchmarks is force of habit.
But that’s a pretty shoddy reason for anybody working in tech,
an industry defined by innovation, to fall back on.
Moreover, I know I’m not alone in thinking that upgrading
to a higher resolution is the primary reason to buy a high-end
PC or graphics card. Enthusiast PC gamers by their nature are
well informed and make sensible purchasing decisions. For
instance, I’ve seen sales data showing that over 80 per cent
of monitors sold with a GeForce RTX 2080 Ti graphics cards
have resolutions of 2,560 x 1,440 or higher. So with a new
year now with us, can we finally get rid of this unsavoury
and aging practice?PCgamersbytheirnature
arewellinformedand
makesensible
purchasingdecisionsUNREALISTIC AND
MISLEADING
James Gorbold is fed up of benchmarketing