File1, 2, & 3 for rburks:



We've been tracking the monthly updates of the Steam Hardware Survey to
create a more detailed window into the GPU replacement cycle than we've
typically provided in the past. There are a number of interesting trends
currently playing out in the GPU market. Nvidia is the midst of an
all-hands-on-deck effort to convince gamers that ray tracing is the Next
Big Thing and that its Turing GPUs represent a worthwhile investment, even
considering their increased costs relative to previous generations. AMD
has aggressively positioned its lower-end GPUs to combat Pascal and
Turing, with a lot of market buzz around 7nm and its upcoming Navi family.
So how are consumers responding to these arguments?

Probably not as well as Nvidia would like, though the company remains the
overwhelming player in the gaming GPU market. According to Steam, Nvidia's
overall market share is around 75 percent, with 10 percent of gamers on
Intel solutions, and 14.7 percent using AMD. There's a little good news
for AMD in these results that we'll discuss as well. First, though, let's
check out the state of Turing versus Pascal.

The slideshow below compares the percentage of Steam users with a given
GPU, measured in the months after that GPU launched. There's a 0 percent
period in the graphs to show the time lag between when cards debut and
when they actually appear in the Steam Hardware Survey. If a GPU launches
in May, May is considered to be Month 1 of launch. The RTX 2080 and 2070
use a 7-month period to reflect the time since launch, while the RTX 2060
uses a three-month window (it launched in January).


I've dropped the 1080 Ti from these comparisons because the Steam Hardware
Survey suffered a major discontinuity in terms of the data set back in
August 2017, and we're now bumping into that period relative to the 1080
Ti's launch window.

Because Turing GPUs sell at higher prices than their Pascal counterparts,
we've also included price-matched comparisons that compare cards based on
their actual price rather than branding. In these cases, the GTX 1080 is
compared against the RTX 2070 and the GTX 1070 takes on the RTX 2060.

So, what do we see in aggregate? Mixed results. The gap between the 1080
and 2080 widened by a fraction, but scarcely enough to notice. The gap
between the 1070 and the 2070, on the other hand, exploded. Adoption of
the GTX 1070 surged once the cards were widely available in-market, while
the RTX 2070 has yet to benefit from an equivalent leap. The GTX 1080
versus RTX 2070 comparison shows improvement, with the RTX 2070 gaining on
the GTX 1080 as far as current adoption at the same place in their
respective life cycles. This is good news for Nvidia.


The RTX 2060 similarly shows mixed results. Steam appears to have a cutoff
at roughly 0.15 percent when it comes to whether a GPU rates being
included on the survey. The RTX 2060 hits this adoption rate more quickly
than any other RTX card, appearing in our survey in the third month
post-launch. As you can see, none of the other Turing GPUs hit this point
until Month 4. It also enters the survey at the highest adoption rate
- 0.27 percent, compared with 0.22 percent for the 2080 and 0.17 percent
for the RTX 2070. Again, this is a sign of increased uptake and better

But while the RTX 2060 has had the best introduction of any Turing GPU
judged on SHS adoption, it doesn't hold a candle to either the original
GTX 1060 or the GTX 1070. The availability of multiple GTX 1060 SKUs
complicates this story, which is another reason why the GTX 1070 may be
the better RTX 2060 comparison. Even here, however, the GTX 1070 is
decisively ahead.

Nvidia has said that Turing drove far more revenue than Pascal during the
early days of launch, and that may be true. Nevertheless, the best public
data source available suggests that Turing has not been as widely adopted
by the gaming community as Pascal was at the same point in its life cycle.

No lines are longer than 80 characters, TYVM. Other specified properties aren't being scored automatically at this time so this is not necessarily good news...