00:00:02
AMD Radeon 960 XT. I'm quite excited
00:00:05
about this one. Well, I'm excited about
00:00:08
the 16 gigabyte model at least, which is
00:00:11
what we have here for review. Needless
00:00:13
to say, as with the 8 GB 5060Ti, we
00:00:17
don't recommend the 8 GB 960 XT. Now,
00:00:23
$350 notionally gets you the 16 gig 960
00:00:27
XT. And that's the one that I'm really
00:00:29
interested in checking out today,
00:00:31
principally because its competition is a
00:00:34
$430 Ti, which offers proportionately
00:00:38
less value than the
00:00:41
$55070. Now, if AMD can provide an
00:00:45
alternative that's just as powerful and
00:00:48
has the required 16 gigs of memory and
00:00:51
delivers the right value, well, I think
00:00:53
that's pretty cool. I think it's a
00:00:55
winner. And when we enter that level of
00:00:57
performance, a curious thing happens. A
00:01:01
$350 GPU is delivering the kind of
00:01:03
horsepower found in Sony's $700
00:01:07
PlayStation 5 Professional. The card I'm
00:01:10
looking at today is the Sapphire Pulse
00:01:12
Edition. And to be frank, I'm glad AMD
00:01:14
sent this one for review as opposed to a
00:01:17
premium factory overclocked version.
00:01:19
It's better representative of the cards
00:01:22
you're likely to get. And I've no real
00:01:24
complaint with this one. Whisper quiet
00:01:25
in operation. One PCIe 8pin input, two
00:01:29
display ports, one HDMI 2.1. Basic to
00:01:33
the point, not overengineered. Exactly
00:01:36
what I like in this market sector.
00:01:38
Specs- wise, much the same is true. You
00:01:40
can think of the 1960 XT as almost
00:01:43
literally half a 9070 XT. 4,096 shaders
00:01:47
across 64 compute units
00:01:49
becomes48 over 32. A
00:01:53
256bit memory interface good for 640 GB
00:01:57
pers of bandwidth becomes 128 bit for
00:02:01
320 GB pers. However, the 9060 XT's TGP
00:02:06
total graphics power is higher than
00:02:08
half. And clock speeds are also higher,
00:02:10
quite a bit higher. Boost clocks and
00:02:13
game clocks, I suppose they're good for
00:02:15
official spec comparisons, but the
00:02:17
realworld output definitely skews
00:02:19
towards the higher boost clock figure.
00:02:22
In fact, let's start by taking a look at
00:02:24
GPU clock speeds, performance, and power
00:02:26
consumption for the 9060 XT. Stacking it
00:02:29
up against the RTX 5060 and 5060Ti are
00:02:32
obvious comparison points bearing in
00:02:35
mind the pricing. But in comparing it to
00:02:37
another AMD product, uh I found the
00:02:40
vintage 2020 6800 XT to be a very
00:02:45
closeerforming GPU to what the Radeon
00:02:48
team has delivered with the 960 XT,
00:02:50
which should give you some idea of the
00:02:53
ballpark output of the card and of the
00:02:56
big big efficiency improvements between
00:02:58
RDNA 2 and RDNA 4. Let's kick off with
00:03:01
some rasterization data where I've
00:03:03
chosen for Horizon 5 because of all the
00:03:06
benchmarks I had, this is where the 960
00:03:08
XT hits its highest clocks over 3.3 GHz,
00:03:13
though the range can vary quite
00:03:14
dramatically between titles. 9060 XT has
00:03:18
90% of the 5060Ti's performance, but
00:03:21
typically requires 25% more power. The
00:03:24
6800 XT is 10% faster, but requires a
00:03:28
lot more juice to get there. RTX 5060
00:03:31
performance is relatively poor here, but
00:03:33
the efficiency differential is
00:03:35
remarkable. Forza has always been
00:03:37
somewhat strange in benchmarking and
00:03:39
there are better results elsewhere. Here
00:03:41
in Alen Wake 2, for example, clocks are
00:03:43
lower, but the performance differentials
00:03:45
are better. 9060 XT, 5060Ti, and 6800 XT
00:03:50
perform much the same. Again, the Nvidia
00:03:52
card uses less power to get there, but
00:03:55
the 6800 XT uses around 120 W more to
00:03:59
deliver much the same output. I'd
00:04:01
imagine results for the 8 gig 960 XT
00:04:04
would be much the same, meaning a
00:04:06
commanding win for AMD over the 5060,
00:04:09
albeit with higher powered draw. Quick
00:04:12
look at RAID tracing, and you'll see
00:04:13
that GPU core clocks drop significantly
00:04:16
with heavy RT workloads. We're now in
00:04:18
2.8 8 GHz territory, while PowerVore is
00:04:22
now in the 180 W range for this Sapphire
00:04:26
Pulse card. With this game, there's
00:04:28
still clear water between the 5060, but
00:04:31
the Nvidia card consumes a lot less
00:04:33
power. Same thing with the RTX 5060Ti
00:04:36
really, though across the whole
00:04:37
benchmark, it can only 5% more
00:04:39
performance. The circa 120 W
00:04:41
differential with 6800 XT remains in
00:04:44
place, but you're getting 20% higher
00:04:47
throughput. Not bad. Bearing in mind
00:04:49
we're talking about 72 RDNA2 compute
00:04:52
units against 32
00:04:54
RDNA4. RT comparisons against Nvidia
00:04:57
will vary dramatically. However, as you
00:04:58
can see here in Alen Wakeake 2, while
00:05:01
960 XT core clocks are higher, the
00:05:03
existing power differentials remain very
00:05:05
similar. But this time, the new AMD
00:05:07
offering gives performance that's pretty
00:05:09
much identical to the nonTi 5060 across
00:05:12
the length of the benchmark. 5060 Ti is
00:05:14
a good 20% ahead while consuming less
00:05:17
power. That said, the less said about
00:05:19
the problem games we've identified with
00:05:21
the Black Core architecture, the better.
00:05:23
But we'll talk about that in more depth
00:05:25
in the benchmarks to follow. But before
00:05:27
we go on to those numbers, regular
00:05:28
viewers may have seen some pretty
00:05:30
alarming benchmarks in my 5060 review.
00:05:33
Essentially, both 5060 and 5060Ti have
00:05:36
PCIe 8X interfaces havinging potential
00:05:39
bandwidth. No problem on a PCIe Gen 4 or
00:05:42
Gen 5 PC, but definitely a problem on
00:05:45
older Gen 3 systems where bandwidth
00:05:47
between GPU and systems suddenly gets a
00:05:50
lot more constricted. I reran those
00:05:52
tests on the 9060 XT again using my
00:05:55
Ryzen 7 9800 X3D using an ASUS X87E
00:06:00
motherboard which can simulate PCIe3
00:06:03
bandwidth. PCIe Gen 3 is slower than Gen
00:06:06
5, but not in any way that should impact
00:06:08
your purchasing decision. Generally
00:06:10
frame rates were the same or a touch
00:06:12
lower, 1 to 2% lower. There is an
00:06:15
exception, however, F-124. Here with
00:06:18
rate tracing maxed, we lose 3%
00:06:20
performance at 1440p and 4% at 1080p.
00:06:23
The result is actually worse with RT
00:06:26
disabled, a 5% differential at 1440p,
00:06:29
rising to 7% at 1080p. So let's talk
00:06:33
numbers then and we have a full face off
00:06:35
between the three 60class cards Nvidia
00:06:37
has out right now and the 9060 XT which
00:06:40
aims to take out all of them. So RTX
00:06:43
5060 and 5060Ti 8 gig up against Ti 16
00:06:48
gig and the new 9060 XT. Looking at Alan
00:06:51
Wake 2 with RA tracing at 1080p
00:06:53
resolution. I have the 5060 Ti here. uh
00:06:57
15 to 16% faster than 9060 XT which is
00:07:01
essentially offering the same kind of
00:07:02
performance as RTX 5060. RA tracing
00:07:05
therefore remains an Nvidia advantage
00:07:08
not least owing to its traditional
00:07:10
hardware advantage along with deep
00:07:12
integration with developers like Remedy.
00:07:15
So a 16point lead for a card that in its
00:07:18
16 GB iteration costs 22% more.
00:07:22
Something to think about. Still, AMD's
00:07:25
RT game has improved significantly with
00:07:28
RDNA4. And where Nvidia is weak, AMD can
00:07:32
deliver a surprise or two, uh, like here
00:07:34
in Avatar Frontiers of Pandora, a game
00:07:37
that traditionally performs poorly on
00:07:39
the Black Row architecture. Here, the
00:07:41
1960 XT is 10.5% to the better against
00:07:45
the 5060, while the Ti lead is 9 to 10%
00:07:49
ahead, depending on the model.
00:07:51
Interestingly, in my tests, certain
00:07:53
games do seem a touch faster on the 8
00:07:55
gig 5060Ti for some reason. Uh, they can
00:07:58
be the same or slower. Here, 8 GB has a
00:08:02
1% lead, which could be margin of error.
00:08:05
Of course, frame times are definitely
00:08:07
spikier on the 8 GB 5060, though weirdly
00:08:10
not on the 8 GB Ti. Elsewhere though,
00:08:14
Nvidia pulls ahead, sometimes
00:08:16
convincingly so. Both 5060 Ti here are
00:08:19
circa 19% ahead in Cyberpunk
00:08:22
2077. Almost but not quite getting to
00:08:25
price versus performance par here. The 8
00:08:28
GB 5060 is on par with the 9060 XT
00:08:32
across the length of the benchmarking
00:08:34
sequence. Dying Light 2 also offers 14
00:08:37
to 16 percentage points advantages to
00:08:40
5060Ti at 1080p resolution on those high
00:08:43
RT settings. And once again, 9060 XT is
00:08:46
effectively on par with 5060 performance
00:08:48
across the length of the sequence tested
00:08:50
here. You'll note that the two jostle
00:08:53
for position across the bench. AMD is
00:08:55
stronger in less dense areas. Nvidia in
00:08:58
more challenging content. Before we move
00:09:00
on from RT though, some notes on 1440p
00:09:03
benches where the 8 GB models struggled.
00:09:06
Both of these cards have issues at the
00:09:09
beginning of our Alam Wakeake 2 test
00:09:10
scene before leveling out. Meanwhile,
00:09:13
the frame time fluctuations seen with
00:09:15
the 5060 in Avatar Frontiers of Pandora
00:09:18
are worse at
00:09:19
1440p, though the 8 GB Ti still looks
00:09:23
pretty okay here. Meanwhile, in some
00:09:25
areas of our Plague Tale Recquum bench,
00:09:27
both the 8 GB cards here can collapse
00:09:30
while both of the 16 gig cards benched
00:09:33
here just crack on. On to rasterization
00:09:36
then where as the tried but tested
00:09:38
cliche always says we should expect to
00:09:41
see a better showing from AMD. So
00:09:43
returning to Alan Wakeake 2, the 5060
00:09:45
Ti's performance advantage basically
00:09:47
disappears. 9060 XT is marginally ahead
00:09:50
of both 8 and 16 GB cards while
00:09:53
delivering a 17point lead over the
00:09:55
vanilla 5060. Elsewhere though, what
00:09:57
you're generally seeing is 9060 XT
00:09:59
clawing back at 5060 Ti's performance
00:10:02
lead, offering superior frames per
00:10:05
dollar in the process. So here in Hitman
00:10:07
World of Assassination, performance
00:10:09
differentials vary between the scenes,
00:10:11
but at the end of the day, this AMD
00:10:13
friendly game offers a two to three
00:10:15
percentage point lead over the TI, 23
00:10:18
points clearer
00:10:20
5060 force Horizon 5. Next, extreme
00:10:23
settings 8x MSIA. It's a pitch battle
00:10:25
across this sequence with the TI cards
00:10:27
delivering a 3 to 5% performance lead.
00:10:30
9060 outpacing the 5060 by 15%. But
00:10:35
let's finish up with uh the Unreal
00:10:38
Engine 5 Juggerauts in our benchmarking
00:10:40
suite. Kicking off with Blackmmith
00:10:42
Wukong. The 5060Ti claws back some
00:10:45
points here with a 12 to 13 point
00:10:47
advantage over the 9060 XT, but the AMD
00:10:51
offering is just 8 to 9 percentage
00:10:52
points ahead of the 5060. Moving on to
00:10:55
Hellblade 2, the 5060Ti's advantage over
00:10:58
the 9060 XT is a mere 3.6% across the
00:11:02
duration of the bench with the 8 GB
00:11:04
cards fairing badly owing to memory
00:11:06
limitations. Interestingly, when a card
00:11:09
is massively overcommitted, the extent
00:11:11
of the impact to performance can vary on
00:11:13
a runby-run basis. In the run you're
00:11:16
seeing here, the 5060 NTI makes it
00:11:19
through this part of the sequence with
00:11:20
no issues. But the Ti has big problems
00:11:23
with kind of strobing frame times there.
00:11:26
Further on, there's a gigantic collapse
00:11:28
to 8 GB performance, but weirdly the
00:11:30
5060 non TI recovers faster than the Ti.
00:11:34
Different runs would likely show
00:11:35
different issues, but the fact is that
00:11:37
even at 1080p, you will find games where
00:11:40
you run into problems on an 8 GB card,
00:11:43
which is why we do not recommend them,
00:11:45
regardless of whether they come from AMD
00:11:47
or Nvidia. Speaking of which, in the
00:11:50
5060 review, I quickly came up with some
00:11:52
VRAMm tests um based on not maxing
00:11:55
everything out in order to prove that
00:11:56
even with reasonable settings and
00:11:58
upscaling, you can still max out 8 GB of
00:12:02
memory. Once again, Marvel Spider-Man 2,
00:12:04
which supports FSR4 and DLSS, proves the
00:12:07
point. 5060 is running artificially slow
00:12:10
here owing to the lack of memory, but
00:12:12
the 16 GB cards have no problems at all.
00:12:15
In this case, the 5060 Ti has a 3%
00:12:18
advantage over 9060 XT. But the overall
00:12:23
point is that yes, budget cards running
00:12:25
with a 1440p output can accommodate raid
00:12:28
tracing. Just don't max everything out.
00:12:31
Moving across to Monster Hunter Wilds
00:12:33
again with realistic settings with high
00:12:35
RT enabled. The 8 gig 5060 keeps pace
00:12:39
for much of the benchmarking sequence
00:12:40
until this happens. Complete collapse as
00:12:43
we run out of memory. The 16 gig cards
00:12:46
are still delivering perfectly
00:12:47
reasonable performance though, so the
00:12:49
compute capabilities can clearly handle
00:12:51
RT, but 8 GB of RAM can't. The reason
00:12:55
I've included the 5068 gig here isn't so
00:12:58
much to show how bad the situation is
00:13:00
for the 5060 specifically, but also to
00:13:03
give you a hint of what could happen if
00:13:05
you do purchase the 8 GB version of the
00:13:08
9060 XT. Now, both of these games we've
00:13:11
just tested support FSR4 upscaling,
00:13:14
meaning you're getting excellent image
00:13:16
quality even at 1440p balanced. Both
00:13:19
support transformer model DLSS, so
00:13:22
Nvidia has excellent image quality, too.
00:13:24
But here's the thing. While FSR support
00:13:26
is gathering pace, Nvidia is always
00:13:29
ahead in terms of supported games going
00:13:31
back to 2020. AMD is seemingly always
00:13:35
playing catch up on features, and that
00:13:37
can't be factored out. FSR4 is a
00:13:39
brilliant first step, but looking at
00:13:41
AMD's Project Redstone, this is
00:13:43
effectively a promise to match Nvidia's
00:13:45
other features that have already
00:13:46
shipped. It's not quite good enough
00:13:48
because you know Nvidia isn't going to
00:13:50
stay still. It's going to continue to
00:13:51
innovate. So support for FSR4 is going
00:13:54
to be key for AMD. But right now there
00:13:57
is actually a tool called Optiscaler
00:14:00
that can actually help retrofit FSR4 to
00:14:03
a bunch more games. Works like this. You
00:14:06
copy its files into the game directory,
00:14:08
run through a quick setup file, and then
00:14:10
in game you can use the inputs from
00:14:12
other upscalers to feed FSR4. And when
00:14:16
it works, it's great. FSR3 on Cyberpunk
00:14:19
2077 looks really poor, but with
00:14:22
Optiscaler, the 1960 XT produces a
00:14:25
superb FSR4 image comparable to Nvidia
00:14:29
DLSS. This is a game with tight Nvidia
00:14:32
integration, though, so 5060 Ti is still
00:14:35
faster. It's ahead by 14 points in the
00:14:37
City streaming test we use. Um 1440p
00:14:40
balanced upscaling and ultra RT. Moving
00:14:43
on to taxing Phantom Liberty content,
00:14:46
the performance differentials remain the
00:14:48
same, but with lower frame rates. Even
00:14:50
so, Ultra RT on this game is no walk in
00:14:53
the park. There's further room for
00:14:55
optimized settings and improving frame
00:14:57
rates. But even here, it's 49 FPS on AMD
00:15:00
playing 56 FPS on Nvidia. Room here for
00:15:04
frame generation then, but that's
00:15:06
another story. Onto further custom
00:15:08
testing now, and I went into this one
00:15:10
with something on my mind. When PS5 Pro
00:15:12
was announced, various press went on a
00:15:15
mission to build a PC with similar specs
00:15:18
at a price point not a million miles
00:15:19
away from the pros $700. Well, back then
00:15:23
I anticipated that a GPU similar to a
00:15:26
downc clocked RTX 4070. It's probably
00:15:29
the closest GPU you could get, but that
00:15:32
was before RDNA4, before the 9060 XT,
00:15:35
and to be fair, before the 5060 Ti. Both
00:15:39
of these cards are pretty good at
00:15:40
achieving prolike performance levels,
00:15:43
sometimes with overhead to spare. Here's
00:15:45
Blackmmith Wukong. Not the most
00:15:47
optimized of console games, but a good
00:15:48
start. Performance mode on Pro still
00:15:50
runs at native 1080p. And while frame
00:15:53
generation is bafflingly used elsewhere,
00:15:56
the prologue stage does not use it. And
00:15:58
as we're running unlocked effectively,
00:16:00
we can compare our PC GPUs at the same
00:16:03
settings to the Pro. Well, I say the
00:16:05
same settings. We've got horsepower to
00:16:07
spare, meaning that I can use DLSS
00:16:09
performance mode, upscaling to 4K on the
00:16:12
5060 Ti, while I'm using the Optiscaler
00:16:15
mod on the 9060 XT to force on FSR4
00:16:19
performance mode. Same internal
00:16:21
resolution as Pro then, but with the
00:16:23
upscaling overhead on top. Even so, 5%
00:16:27
faster on 9060 XT versus Pro, rising to
00:16:30
27 points clear on 5060 Ti. One might
00:16:34
imagine that DLSS overhead is lower than
00:16:36
FSR4 here. Oh, and another change in
00:16:39
this comparison. Uh, the Pro version
00:16:42
mysteriously uses low quality textures.
00:16:45
I bumped that on to high when testing
00:16:47
the 9060 XT and 5060 Ti. More Pro
00:16:51
comparisons. Well, here in Fortza
00:16:52
Horizon 5, we're capped at 60 FPS on the
00:16:55
console. This version of the game has a
00:16:58
60fps performance mode that takes the
00:17:00
existing base console configuration and
00:17:03
spruces up environment settings, which
00:17:05
I've matched here. If it's not a
00:17:07
complete match, it's very, very close.
00:17:09
We can't see the full power of the Pro
00:17:11
here with that VSYNC cap, but both 9060
00:17:13
XT and 5060Ti average at over 70 fps
00:17:17
with a 2.7% lead for the 5060 Ti over
00:17:20
9060 XT. Native 4K on a so-called budget
00:17:25
card. Well, this is a game built for 4K
00:17:27
output on consoles, so it's no surprise
00:17:29
to see today's mainstream GPUs able to
00:17:32
do the same. Let's look at Alan Wakeake
00:17:35
2 next. We'll begin with a comparison to
00:17:37
PS5 Pro's performance mode. This
00:17:39
upscales from 864p to 4K using FSR2 with
00:17:44
quality settings equivalent to the base
00:17:45
conso's quality mode, which is
00:17:48
essentially PC's medium. This is easy to
00:17:50
replicate then and as the Pro is mostly
00:17:52
under 60 frames pers, we are finding the
00:17:54
limits of the GPU. 9060 XT 13% faster
00:17:59
than Pro 5060Ti is 10% ahead. So
00:18:02
certainly in rasterization terms, um
00:18:05
these cards are very close to what
00:18:07
Sony's $700 console is doing. But what
00:18:10
about RA tracing? Alan Wake 2. Next up,
00:18:12
we're going to be taking a look at the
00:18:14
quality mode uh capped to 30 frames
00:18:17
pers. Now here's the problem. The RT
00:18:19
setting on PS5 Pro is effectively lower
00:18:22
than PC's low. Even so, I thought I'd
00:18:25
give it a go. 4K output resolution using
00:18:27
FSR2 balance mode, effectively 1270p
00:18:31
internal. Then, as it is capped at 30,
00:18:33
we aren't seeing the pros full power.
00:18:35
But if we freeze frame here, we're
00:18:38
dipping beneath 30fps on the console.
00:18:40
So, we have found pro limits. You'll see
00:18:42
that both of our PC alternatives
00:18:44
struggle. We need that lower than low RT
00:18:47
setup to get a true measure of how close
00:18:50
we can get to the console experience.
00:18:52
Even so, I think we've proven out that
00:18:54
we're basically in the ballpark of PS5
00:18:57
Pro GPU performance across the board.
00:18:59
And with the 1960 XT, that performance
00:19:02
won't dwindle that much if you're
00:19:04
slotting the card into a PCIe gen 3based
00:19:07
PC. For $350, that's a very good deal,
00:19:10
even if the days of stratospherically
00:19:12
good genon GPU performance increases in
00:19:15
the mainstream market are long gone. So,
00:19:18
in summary, here's how I feel about the
00:19:20
1960 XT. Kind of slots into a market
00:19:23
defined by three Nvidia GPUs, only one
00:19:26
of which has enough memory in my
00:19:28
opinion. Well, based on my tests, you're
00:19:30
getting the lion's share of 5060Ti 16 GB
00:19:34
performance while being cheaper than the
00:19:35
8 GB model. assuming MSRP is real, of
00:19:39
course. Um, RT performance not as strong
00:19:42
as 5060Ti overall, but by and large,
00:19:45
Nvidia's performance isn't in line uh
00:19:48
with its higher price. 9060 XT clearly
00:19:51
offers better value. What Nvidia does
00:19:54
have is far far superior support for
00:19:56
upscaling, and Radeon users shouldn't
00:19:59
have to install mods to get FSR4 running
00:20:02
on more games. At the very least, FSR4
00:20:05
needs to hit FSR2 levels of adoption.
00:20:08
Frame gen support is also disappointing
00:20:10
right now on the AMD side. And while its
00:20:12
effectiveness lessens on budget level
00:20:15
cards, frame generation, multi-frame
00:20:17
generation, it does have utility,
00:20:20
especially when there are so many high
00:20:22
refresh rate monitors on the market
00:20:24
right now. I mean, this is Cyberpunk
00:20:27
2077 1440p display output, full RT
00:20:32
overdrive running on the RTX
00:20:35
5060Ti. Fay rates well north of 100
00:20:38
frames pers. Pretty decent high refresh
00:20:41
rate experience. Actually, the only
00:20:43
slight downer being that latency is a
00:20:45
bit on the high side, but perfectly fine
00:20:48
actually when using a controller. And I
00:20:50
think it would be remiss of me not to
00:20:52
point out that as good as the 9060 XT
00:20:55
is, sometimes those Nvidia features just
00:20:58
come together and produce something like
00:20:59
this. Pathfraced Cyberpunk 2077 on a
00:21:04
60class card. That is pretty special,
00:21:07
but it doesn't suddenly make the value
00:21:09
problems facing the 5060 Ti go away,
00:21:12
right? So, it's a tricky one, but
00:21:13
Nvidia's problem does diminish if it
00:21:16
gives the Ti model the price point it
00:21:18
actually deserves. The 16 GB model
00:21:21
shouldn't be more than $400 in my
00:21:23
opinion, max. And even without the
00:21:25
arrival of 9060 XT, the 5070 offers
00:21:29
proportionately more value than that 16
00:21:31
gig 5060 Ti. That's not what budget
00:21:35
gaming is all about. I've said it many
00:21:37
times, but price versus performance
00:21:39
should improve as you move down the
00:21:41
stack into the mainstream sector, not
00:21:44
get worse. We've crunched the numbers,
00:21:46
and here's how dollars per frame works
00:21:48
out. RTX 5068 GB is on top, but we'd
00:21:52
have to rule that out as a
00:21:53
recommendation simply because while our
00:21:55
benchmark games work fine at 1080p
00:21:58
overall, we've proven that AAA is a
00:22:00
problem for 8 GB cards going forward.
00:22:04
9060 XT 16 gigabytes is next, but
00:22:07
remarkably its RDNA4 stablemates follow
00:22:10
before we hit the
00:22:12
5060Ti. The 8 GB version of the 5060Ti
00:22:16
model won't have the longevity we expect
00:22:19
from the card. So, we need to factor
00:22:21
that out. And there we go. 5070 is
00:22:24
offering better value than the 16 GB TI.
00:22:28
That needs to change. And here's that
00:22:30
same table at 1440p resolution. 9070 and
00:22:33
9070 XT deliver proportionately better
00:22:36
value followed by 5060 and 5070. Dodgy
00:22:40
pricing and diminishing performance
00:22:42
combined to show that the TI cards place
00:22:45
beneath that. Again, good evidence that
00:22:48
a price adjustment is needed. Of course,
00:22:50
do bear in mind that these are our
00:22:51
benchmarks and our mix of Raster and RT.
00:22:55
Other outlets will have a different mix
00:22:57
and different games and therefore
00:22:59
different results. But that's all for me
00:23:00
on this one, which means I can revert to
00:23:02
the standard appeals for likes, shares,
00:23:05
subscriptions, bell ringing, whatever.
00:23:07
And yes, please do consider backing all
00:23:09
of our work via the DF supporter
00:23:11
program. High-quality video downloads of
00:23:13
everything we do, weekly news updates
00:23:15
from the team, a team you can actually
00:23:17
interact with on our Discord, early
00:23:19
access to a bunch of stuff, including
00:23:21
DFDirect weekly. Oh, and occasional
00:23:24
bonus material, too. Finally, please
00:23:26
check out store.digitalfoundry.net net
00:23:29
for our merchandising wares. Hope you
00:23:31
enjoyed this one in the meantime, but
00:23:33
for now, thanks for watching and
00:23:34
supporting Digital Foundry.
