• 17 hours ago
Finally, AMDs new 9070 and 9070XT vs the competition! Now that all the GPUs are finally out on the table (literally) who do you think won the 2025 GPU war?

Sponsored Links:
New Users sign up for a Free World of Warships Account and receive the following! 7 Days Premium Time
300 Doubloons
1,000,000 Credits
Tier V Ship: HMS Exeter!
Claim your new player rewards here! - https://wo.ws/3Z5lIWa

○ Get your JayzTwoCents Merch Here! - https://www.jayztwocents.com
○ Join this channel to get access to perks:
https://www.youtube.com/channel/UCkWQ0gDrqOCarmUKmppD7GQ/join
○ Join the official JTC Discord! https://discord.gg/jayztwocents

○○○○○○ Items featured in this video available at Amazon ○○○○○○
► Amazon US - http://bit.ly/1meybOF
► Amazon UK - http://amzn.to/Zx813L
► Amazon Canada - http://amzn.to/1tl6vc6

••• Follow me on your favorite Social Media! •••
Facebook: http://www.facebook.com/jayztwocents
Twitter: http://www.twitter.com/jayztwocents
Instagram: http://instagram.com/jayztwocents#
SUBSCRIBE! http://bit.ly/sub2JayzTwoCents
Transcript
00:00Thank God
00:01This is like the last launch for a little while when it comes to GPUs
00:05What you see on this table is not even a third of all the GPUs. We've retested in the last month and a half alone
00:13Anyway, this is a twofer. You're getting 9070 XT and
00:199070 from AMD in one video because there's just no point in splitting them into two
00:30Hey day day
00:40What we got work to do yeah, I'm playing World of Warships
00:47World of Warships is the free-to-play naval strategy game where you command the most iconic and famous warships from World War one and World War
00:53Two recreated with stunning detail and accuracy build your fleet while participating in various game types while upgrading your ship's arsenal along the way
01:00New players who sign up using my link below will receive an exclusive starter pack to get you up and running quickly by receiving seven
01:06Days premium time 1 million credits 300 doubloons and the tier 5 premium ship the Exeter
01:11So what are you guys waiting for? Start sinking ships with World of Warships by heading to the description below and getting your freebies
01:19Okay, so 9070 and 9070 XT it's interesting this time around because as you can see by some of these other reference cards on the
01:26table from AMD
01:28There is no reference card for the 9070 series cards from AMD this time around
01:35So all of the AIB partners kind of have carte blanche to do whatever they want
01:39In terms of the designs and such. In fact, even the reference spec is almost more or less a suggestion
01:46We'll talk about that a minute. So if you haven't seen one of our reviews before thanks for coming by. I'm JST sense
01:51We're gonna talk about performance power draw temperature generational
01:53actually, you know, we're not gonna talk about generational performance uplift because this is technically I don't even know exactly which card it would
02:01Compare to if you want to know the truth
02:03because the previous naming scheme was like 7900 XT XT X and then 7800 and then
02:117700 and the 9070 does not tear up with a
02:1677 so the sevens don't match if you will
02:19So essentially all we're gonna do today is just here's where it landed on the charts and you can see how depending on the title
02:24It's very it very much impacts which cards it sort of lands against now cards
02:29You're gonna want to kind of pay attention to on the charts as we talk about them
02:32It's gonna be like the 7900 XT the 7900 XT X the 5070 TI from NVIDIA
02:39Which is very close in terms of price parity to the 9070 XT, but then obviously the 5070
02:46When we compare it to the 9070 you see why they named it that way and they did that on purpose
02:50They even admitted it. So that's at least that sort of makes sense
02:54Anyway, 5070 FE on our chart, which is an MSRP card versus the 9070 Quicksilver here, which is not an MSRP card
03:02So let's talk about the cards that we have real quick
03:04We've actually received a few of them
03:05But the only two that have like finished validation and testing for us is the XFX variants here
03:10Which I'm kind of okay with even though they're gonna be some of the more expensive models on the market
03:15The white one right here. This is our 9070 XT
03:19that is the
03:20Mercury or Merc and it is the magnetic air version and I was like what the heck is magnetic air when I first heard about
03:26It well as you can see you can just pop the fans off
03:29They're just magnetic actually and then the pogo pins is contact to a 360 pad that's on the GPU
03:35You might be asking yourself. Why would anyone want to do this?
03:38Well, I honestly firmly believe it to be just the ease of cleaning your fans getting it out getting more the dust out
03:46It's actually a nice feature to have also to it
03:48if they continue to
03:49manufacture these and make parts and replacement parts for them if you have a fan that starts to wobble or a bearing that starts to
03:54Go bad or any sort of fan
03:56Chatter noise that happens over time. You can replace the fan. So I've never seen a graphics card do this before
04:02I'm not saying it hasn't existed. I just personally haven't seen or touched one. It's a pretty neat feature
04:07In fact, it's pretty nice to wear it because it's magnetic
04:09It literally just indexes and puts itself work back where it wants to it's a very strong
04:14magnet this particular card is also a higher than reference spec what I mean by that is the
04:209070 XT is a 304 watt TVP not 300 not 305
04:27304 I mean that is oddly specific
04:31anyway, the 9070 is a
04:33220 watt card, but I point this out because the AI B's are actually allowed to go all the way up to 340 watts
04:40On their out-of-the-box specs and we do have a 10% additional power limit available to us as well
04:46But all the testing was done out-of-the-box settings with fan curves and power limits
04:50No rage mode or any of that stuff just out-of-the-box on the test bench install the drivers and off we go. No tuning
04:56No, nothing. Okay, so key specifications here 9070 both cards are obviously on RDNA 4. We've got
05:033584 stream processors the 9070 XT has
05:074096 stream processors
05:09So obviously 512 more that's that's how numbers in arithmetic work
05:14Anyway, when it comes to RT accelerators are the same as RT cores if you will
05:1956 on the 9070 and 64 on the 9070 XT and then AI processors or like tensor cores
05:25Which is for like upscaling and such you have 112 on the 9070 versus 128 and the 9070 XT and both of them are listed
05:32at having 128
05:34ROPs
05:36I'll admit I didn't actually verify that all 128 are there
05:40Sucks that I have to do that
05:42But what were the performances landing? I'm gonna assume they're all there
05:46That sucks that I forgot to do that, but just being fully transparent with you guys
05:49I'm not in the habit of testing the fact that we're are all the cores and stuff there that were promised
05:54Thanks Nvidia when it comes to frequencies, though
05:57The 9070 has a 2070 megahertz game clock and a 2520 megahertz boost clock on a 220 watt TVP
06:03Recommending a 550 watt PSU or higher 9070 XT is a 2400 megahertz game clock
06:112970 megahertz boost clock with that 304 watt TVP recommending a 700 watt PSU
06:17But since most cards are gonna actually probably be about 340 watts
06:21It's probably recommended to just move that number to 750 watts when it comes to recommended PSU
06:25Now when it comes to pricing this is where I say
06:27It's a little bit of like Wild West slash carte blanche when it comes to the AIBs
06:32SAP are suggested detail price, which is recommended by an almost at Nvidia AMD is set at
06:39$549 for the non XT and then the XT being at
06:43$599 only $50 of variance between the cards most people in the announcement video felt that that was too close together
06:50Given where a lot of the spec differences are between the two cards spec wise
06:55They're actually very similar as you saw both cards are featuring 16 gigabytes of GDDR6
07:01In fact, they're the exact same memory configuration on both cards
07:04I want to point that out like neither one gets any sort of a speed bump or a spec bump when it comes to the
07:09Memory, they're both 256 bit memory buses with a throughput of six hundred and forty four point six gigabytes per second
07:16Exactly the same on the memory
07:17so the only difference is like I said as core clock and then 512 additional stream processors found on the XT and then obviously a
07:25handful of more AI and
07:28Raytracing cores that are on there. So
07:31The prices are going to be kind of all over the place though because cards can play around with the 40 watt
07:35Well 36 watt additional power limit available to it the white card that I have right here
07:41Which is the like I said the magnetic air 9070 XT from XFX the mercury card comes in at seven hundred and sixty nine
07:47Dollar MSRP. This is the first time in a review though. I've actually had a brand tell me what the
07:53worldwide
07:54SEP is versus the United States
07:57Manufacturers suggest retail price because of either tariff or no tariff. So
08:02769 for the white 749 for the black. It's $20 more for the white. So that puts it at about let's just say black
08:08I don't I don't think 20 extra bucks for white is obviously worth it unless you're building a white build
08:14So let's just say 749. Okay for pricing that makes it one
08:19$150 more expensive than
08:21SCP if you compare like the NVIDIA 5070 TI at a
08:28$749 MSRP but most cards are actually up over
08:31$900 that's actually a closer to SCP price for one of the most custom cards on the market
08:37Which actually is a nice thing that that's kind of sucky to have it cost so much more
08:42But there are obviously other versions that exist like the quick silver
08:45The black we have here is a five hundred and eighty nine dollar and ninety nine cent worldwide
08:50MSRP so that's making it about
08:52What 40 bucks above MSRP that's about the kind of stuff we normally expect to see with so that's a custom card believe it
08:59Or not versus the reference design supplied by AMD
09:02So that's a more reasonable like cost lift for a custom model with a giant cooler on here
09:09Anyway, when it comes to the estimated shelf price after tariff, which is the first time I've ever had this figure
09:16Eight hundred and forty nine dollars and ninety nine cents for this white one right here and
09:21$649 and 99 cents for the 5070 quick silver
09:26damn it and
09:28$649 and 99 cents for the 9070 quick silver right here. So
09:34There's that when it comes to pricing so
09:38Take that as you will
09:41Anyway
09:42AMD does say there'll be plenty on shelves. I guess we'll find out tomorrow
09:47Moving on to performance because none of that pricing matters if you don't know how it performs
09:52So in Speedway, which is an RT test
09:54Remember, we've got third-gen RT cores now on 9070 versus second-gen RT cores found on 7000 series the 9070 quick silver comes in
10:03just ahead of the 4070 TI tough and
10:07Literally 13 points behind the 5070
10:09So not bad at all when it when it comes to the performance there the mercury coming in ahead of the 4070 TI super tough
10:17Not a lot but ahead and a fair ways behind the RTX 4080
10:22Not surprising when it comes to a pure RT test
10:25There is an entire generation gap between the two now if we move on to rasterization performance in our synthetic here in times by extreme
10:32The quick silver comes in ahead
10:34By 30 points of the 7900 XT at a 12,000 987 and
10:41losing to the 5070 TI prime by a few percent and then the mercury coming in way ahead of the prime at a
10:5114,703 even ahead of the 4080 super and pretty much matching the 7900 XT X don't forget that was a thousand dollar card
10:59this is a
11:01$700 or $600 card 599 so
11:04Except that one's a little bit more obviously
11:07Moving on to f1 and 1440 all of my gaming charts. We're gonna talk about personally are in 1440
11:12I think that's where the card is an absolute beast
11:15Obviously the charts that we will show that you guys can look at as you pause and go forward
11:21What have the other resolutions here?
11:22but our discussion here is gonna talk about 1440 9070 quick silver coming in ahead of the 4070 TI and
11:28Landing just behind the 4070 TI super tough such a stupid name
11:32Anyway moving on the 9070 XT mercury coming in ahead of the 5070 TI prime by 3 FPS and
11:40Losing to the 4080 super by 3 FPS not bad if you consider the fact that this card
11:46Technically its MSRP is only 70% of the MSRP of the 4080 super
11:53Although remember this card is a custom model. So it does carry a
11:56Premium for that moving on to the next title here. Let's talk about black myth
12:01Wukong black myth Wukong 1440 P. This is a title that does not favor AMD
12:06I mean you can see that by the throttle the bottom three cards being a 7800 XT a 7900 XT and a 7900 XT X
12:13Something about the RDNA 3 architecture and black myth Wukong do not get along
12:18however, RDNA 4 has entered the chat and you can see there is a a
12:23bit of an uplift in performance versus previous gen for sure the 9070 quick silver is
12:29Just edging out the 3080 and
12:32Losing to the 3090 by a little bit
12:34That sounds pretty disappointing when you'll see what this these cards do to those particular cards of 30 series in other games
12:40It's just something about black myth does not like AMD cards
12:45But it's nice to see this huge gain between RDNA 3 and 4
12:49When it comes to this title and then the 9070 XT mercury about three and a half FPS behind the 4070 FE
12:57But about 2 FPS ahead of the RTX 3090 so not bad right there
13:03If we're going to talk about Borderlands 3 badass settings here, we know Borderlands 3 loves AMD cards
13:09That's why there should be no surprise that the XFX 9070 quick silver
13:14comes in right in the middle of the chart ahead of the 4070 TI super tough and just behind the
13:207900 XT and then the mercury is actually ahead of the RTX 4080 and
13:27Just behind the 4080 super so not bad at all when it comes to that generational uplift
13:32Although 7900 XT X is still king when it comes to Borderlands 3
13:37avatar frontiers of Pandora
13:39Another title that doesn't really seem to favor AMD a whole lot
13:43But that's why we have our suite that we do because it does really show the strengths and weaknesses of certain
13:48architectures in certain titles
13:49so the 9070 quick silver is way well ahead of the
13:544070 and just behind the 3090 the mercury
13:57However is just ahead of the 90 or 7900 XT and just behind the 4070 TI
14:03Tough you can see that 4070 TI tough. It seems to trade blows with quite a bit
14:08It kind of makes me wonder when they said they were
14:11naming these
14:134070 series cards so that it was less confused confusing for consumers to figure out what card it's designed to really compete with on team
14:20green you can see 4070 was definitely a card that they were
14:25checking out when they were kind of testing stuff out because of the fact that
14:295070 wasn't out yet when they announced all this sort of stuff
14:32But if you pay attention to to where the 5070 is on these charts and the 5070 TI
14:36You can see it's still holding its own pretty well against those cards
14:39Mostly because Nvidia did not give a huge generational uplift that we're used to seeing so it actually fared well for AMD
14:45For Nvidia to have such a lackluster uplift from generational improvements. So I guess that was sort of fortunate for them
14:52anyway, Horizon Forbidden West a title that is heavily favoring AMD and we have seen this across the board the
15:009070 quick silver is just behind the 5070 TI prime and a fair amount ahead of the 7900 XT and the
15:089070 XT Mercury is literally exactly the same 0.1 FPS
15:12And we know that 0.1 FPS is sometimes the make-or-break of whether or not a game is playable
15:17But it is 0.1 FPS behind the RTX 5080 FE in Forbidden West
15:23That I that was unexpected and it is a fair amount a bit ahead of the 4080 super now the quick silver lands
15:29Just under the 5070 TI prime and just ahead of the 7900 XT
15:35The the engine that is used in horizon
15:38Definitely likes AMD cards cyberpunk. This is gonna be a good one. We start our tests up to have a mixture of RT
15:44No upscaling so it's no RT reflections either. It's all lighting global illumination shadows that sort of stuff
15:51No actual reflections and then we do not do an upscaler
15:54So it's really a horsepower test between the like RT cores and then the CUDA core stream processors to figure out the generational or
16:02Comparative uplifts versus other cards. So a pretty decent jump for the 70 series cards coming from AMD
16:09So the XFX 9070 quick silver is above the middle half of the chart
16:13Coming in just behind the 4070 TI super tough by less than 1 FPS behind
16:19and just ahead of the 4070 TI tough
16:22so literally just splitting the hairs right there and then the
16:26Mercury card
16:28Coming in ahead of the 4070 TI tough and just behind the 4080 FE
16:32So in the top half of the card or the chart for AMD for the first time to be honest
16:39When it comes yeah, if you look at the 7900 XT X it's just under the bottom half of the card
16:43So that shows you generationally when it comes to like the RT or the rate the ray acceleration ray tracing accelerators and the AI
16:49Accelerators it bumped it up into the second half of the chart, which is good to see
16:53Stalker 2 last title we'll talk about here
16:57Kind of nothing exciting to talk about. It's just ahead of 70 the 9070 quick silver is just ahead of 7900 XT
17:03just behind 4070 TI super tough and then the Mercury XT is just ahead of the 7900 XT X and
17:12About 6 FPS behind the 5070 TI prime stalker definitely does favor in video graphics cards
17:19But there's more to talk about here than just average FPS
17:22You'll notice if you look at these charts closely the 1% and 5% lows are higher than Nvidia's parody cards
17:30price parody cards on average now, I started doing some digging for that and to try and figure out why and
17:38There's quite the story to tell here
17:39If we look at f1 2024 1440p ultra high preset with ray tracing and no upscaling if we look at frame times alone
17:47You can see here the yellow line, which is
17:50Pretty much covered by the teal line, which is the 5070 TI showing that an average frame time
17:56At least the median if you draw an average line through the frame times
17:59For those two cards are nearly identical in f1 and the FPS sort of showed that but if you'll notice
18:05These spikes and peaks that are in the teal line
18:08The 5070 TI shows huge fluctuation between that average frame time and then the transient spikes or or lows to be honest
18:17I would even so the lows are the lower the lowest frame time to meet back to the average you have highs and lows
18:24Those differences right there can perceive be perceived as jitter stutter
18:29Hesitation pauses depending on how long that spike stays. I like it takes to draw that one frame
18:35So if you look at even the pink line
18:38Which is the 9070 non XT. You can see there's a lot less of those
18:43In fact, if you want to know what really consistent frame times look like we'll take the NVIDIA 5070 TI
18:50Prime from Asus off the chart and this is what you get
18:54You can literally count the spikes like the big spikes
18:58by hand, there's two for the 9070 and
19:02Technically, there's three for the 9070 XT, but the ones at the start of the game
19:06That's f1 when the when the race actually starts. There's a spike on every card right there
19:11There's something to do with the game and the way it runs. So ignoring those is one per card and they're very very consistent. So
19:19Having even those lows come up to the average is better
19:24than having a whole bunch of lows compared to your average because again the difference between the lowest or the fastest time to draw a
19:30Frame in the average that distance is the same as having it above the average
19:35But if you take away the average now and you just compare that gap
19:39That's now the perceived stutter and frame time jitters and stuff that people report when they talk about smoothness of gaming
19:45So I have to say this
19:47there has been an
19:50extremely noticeable amount of smoothness with this generation of AMD graphics card
19:55I also have to say
19:58this is the first and
20:00This is the first time I have had an AMD review with more than one card even one card or not where I didn't have
20:08Some weird game that crashed or some weird game that wouldn't launch for some reason or have to reach out to AMD for help on
20:14why I'm getting some anomalous behavior in some certain title or
20:19So the game just refusing to launch entirely like this is the first time I have had some of the most
20:27Pleasant review experience of a graphics card and this is not just AMD
20:32This is including Nvidia and I'm not just including this generation which we know for it for Nvidia has been a shitshow
20:39I'm going back to the beginning of starting YouTube. I have never had such a smooth
20:45Experience when it comes to AMD graphics cards now
20:49I wanted to make sure that this wasn't just like a cherry-picked title like oh
20:52F124 is just super favoring the Nvidia or the AMD cards and for some reason having all these transient spikes and dips on the AMD
21:00Card, so I went ahead and loaded up cyberpunk 2077 in 1440p. So all the same settings we do our benchmarks with in fact
21:06These are the benchmark runs right here that you're seeing by the way. This isn't a separate test
21:10These are the actual benchmark runs that I've turned into the frame time charts for you guys
21:15But this time I just did 90 70 XT versus 50 70 TI
21:18But again with the exception of like right in the middle there that one really big spike which could be
21:23Anything those types of transient spikes happen in every game. It could be something started in the background of the of the test bench
21:29Just about anything could have caused that and it's really hard to figure out what causes one transient spike
21:34It's when it happens all the time that you have a problem. You got to figure it out, but I want you to notice that
21:39even
21:40The Nvidia card that has a lower average frame rate still has a lot of fluctuation on the lows as well as some of the highs
21:46So myself personally I would prefer to have a more squished average frame time
21:52That's slightly higher than another graphics card that might have a median line drawn through their frame time
21:57That's lower over the bigger range of variance because that range of variances. Like I've said, I'm a broken record
22:04That is what proceed gives you the perceived stutter enos or inconsistency in the frame rate the graphics card
22:10the 90 70 XT in the 90 70 truly felt
22:13like it was a higher frame rate than it was and that's because of the higher 95th and
22:19First or 99th percentile or first first 1% low and 5% low frame type rates
22:25Being higher if you look at the chart
22:28Also to one other thing to note. I don't have a way to really present this to you
22:32I I don't want to have this turn in to be 50 different charts per review
22:37It is the most consistent
22:39The 90 70 XT specifically too is the most consistent run to run
22:44Consistency in terms of the frame rates that we see when we do our testing out of all of the cards
22:50It's not uncommon to see one two, three, sometimes maybe even four
22:55FPS variation run to run or to have one card have a better fit percentile or one run have a better fit percentile than the
23:01Other and then another one has a better first percentile and they just kind of like whack-a-mole which one has the worst one
23:06I'm telling you right now. The AMD card was like within one FPS every single time
23:12In fact so consistent
23:14I found myself actually running the tests
23:16More than our typical three runs and some of them in the beginning when I was validating anywhere between six to seven runs
23:23Just because I was concerned that for some reason I was
23:26Writing the same number and not actually opening the correct test
23:29So I would throw them away and start over and get the same consistency
23:34Sometimes only the difference is being noticeable in the tenth of a frame which we do measure down to we measure down to the tenth
23:40of a frame
23:42It is consistent. It is actually amazingly consistent in that arena. Anyway, here's one other title. I figured okay f1
23:48We know it does really well cyberpunk
23:50We know that that's a there's an RT discrepancy
23:53What about a title that we know doesn't like AMD so black myth Wukong is back up
23:58Here's the frame times between the 9070 XT Merc and the 5070 Ti Prime
24:03You can see right here
24:05If you again, there's only a few of them
24:07But if you compare although it is a a fair amount of average frame time
24:12Which also mathematically turns into frame rate is lower and more consistent like straight line across the board
24:20For the 5070 Ti. I want you to notice though. There is actually
24:25Less transient spikes on the 9070 XT Merc. Okay. So last but not least. It's going to talk about power draw and
24:32Temperatures and frequencies so in Furmark at 4k, which is a very difficult stress test the XFX 9070 XT
24:39Is pulling about 350 watts across the board through our hardware measurement device. That's including PCI Express
24:45Although it's not really pulling any power from PCI Express
24:48That's not surprising considering I told you it's a 340 watt card and there's some heat and efficiency loss through
24:55The operations of making frames happen. So that is what the actual power draw is from the PSU
25:01350 watts so it's it's a bit of a
25:05Increase over say the 5070 Ti. So if you are like the ultra cons like I want to conserve all power at all costs
25:13The obviously just over 300 watt more or less 310 watt
25:185070 Ti Prime is going to be the more efficient card and then
25:22Unless you're like, you know what screw that I want to save you the more power the 9070 Quicksilver comes in at about
25:29228 watts drawn from the wall or from the power supply directly now that's in Furmark
25:34Which is a less than realistic scenario when it comes to gaming obviously games fluctuate a lot
25:39So when we move over to Cyberpunk 1440p
25:42You can see the 9070 XT still draws 350 plus watts. It's about 352 watts on average
25:49Actually, it's pretty consistent for all three of these cards pulling the same amount of power between Furmark and
25:55Gaming so this is one of those few times
25:57We actually see the the figures just carry over to the other chart
26:00Usually we'll see some sort of a drop but that's on the higher-end cards to be honest
26:04But you don't have to go like full tilt to give you their max FPS
26:06Depending on whether or not there's any sort of CPU bottleneck or it's fully GPU bound
26:11but again, just over 300 watts for the 5070 Ti Prime and then hovering around
26:18225 to 300 or 230 watts for the 9070 Quicksilver
26:23So it's a big power discrepancy between the not XT and the XT
26:29Sorry about temperatures real quick in that same Furmark test
26:32There's something interesting with the temperature sensor in the 9070 and 9070 XT. It doesn't seem to update quite as often
26:39You'll notice if you look at the teal line here, which is the 5070 Ti Prime. There's lots of input
26:44There's lots of data points that will give you a very jiggly line, right?
26:50That's because of how quick the card is responding back to the polling software saying what's your temperature?
26:54The 9070 as you can see is slightly warmer than the 9070 XT. Although the cooler is like a very similar size
27:00It's just not nearly as dense to be honest. I have no idea what the thermal interface material is or any of that stuff
27:07I'm not doing a teardown on these cards, but even though it's a much lower power card
27:11it is warmer than the 9070 XT averaging about
27:1553 C I say warmer and 53 which is funny
27:18But that's because the 70 the 9070 XT is reporting a max temperature of 47 degrees
27:25Now this cooler is as big as a 5090
27:30That's why the 5090 is on the table. It is a bit overkill
27:36600 watts
27:37540
27:39600 watts
27:41340 I'm inclined to believe it's actually 47 degrees
27:44I have not put any sort of temperature probes on there to verify that
27:47But I'm inclined to believe it just based on the weight and density of that card alone
27:51Which is why it is coming in at a max of 47 degrees in gaming exactly the same story
27:56but again
27:57You can see what I mean about the update
27:58The amount of data points you get with the temperature sensor on the Nvidia cards versus this family of card
28:04That's why it just it's like
28:05Boop-boop like a little heartbeat and then flatten a little heartbeat
28:08But that is the way that the sensor reports when we export the CSV data. So take that information as you will now before we move
28:16On to frequency here. I want to talk about the fan designs here
28:19I will say this the 9070 XT and the mag fans that are found on here
28:25Actually have a fairly pleasant sound by pleasant
28:28I mean
28:29They sound like bigger fans than they are to me
28:31It sounded like I had an air purifier sitting on my desk when it was at full speed the 9070
28:36Quick silver fans definitely had a more turbulent kind of a choppy weird sound to them
28:41like almost like the blades were moving in some way like it had a
28:45I don't know how to describe it
28:47We have some sound profiles here that you can hear at both 150%
28:50But I do I do feel like the XT with the mag air fans are the magnetic fans or magnetic air
28:55Weird thing to call it sounded better
29:26So
29:28Frequencies if we look at f1 20 24 14 40 P again RT high
29:33You can see that we're definitely fluctuating quite a bit on the 9070 XT, but between 2800 and 2900 megahertz
29:41I didn't put the Nvidia cards on here because the architectures are so different
29:451 megahertz on one is not equal to 1 megahertz on another so just be comparing apples and oranges
29:50So I just wanted to compare the 9070 and 9070 XT versus the numbers that they say are the base numbers
29:56so as we said on here the 9070
29:59has a
30:012520 megahertz boost clock and we were actually landing just hovering over the line of 2500
30:07So we were actually getting the full boost clock the game clock. I hate the fact they call it game clock
30:11I just ignore game clock. It goes to boost clock
30:14Anyway, don't waste your time worrying about what game clock is to be honest and the 9070 XT
30:18Has a 2900 and 70 megahertz boost clock
30:22Did not really ever hit that it actually barely crossed 97 or 2900 at all
30:29It landed more or less around 2800 and 70 megahertz, but again, that is the boost clock
30:34That is the max up to and I'm they were hitting it was hitting power limit long before that
30:38I think if I put the additional 10% slider, I'd probably be able to hit that but with the game clock being 2400 megahertz
30:43We were significantly above that over 2800 the entire time
30:47So kind of surprised at how cool it ran that team
30:50That means we have to be power limited on what's allowing that frequency to not quite hit that max up to
30:55but if there's one thing we've ever learned about AMD cards is
30:58They definitely lean into the up to figure when it comes to all of their like clocks and stuff
31:03All right, final thoughts now that we have everybody's cards on the table
31:07The 9070 9070 XT are very strong competitors
31:12the the the old adage of
31:15AMD is good, but it's got bad drivers, whatever. It's got I
31:22Guess just insert whatever complaint you have here. I've heard them all. I
31:26Don't think they hold up this time around. I think what's happened here is Nvidia has made such terrible
31:34Backward strides in things like reliability
31:38mostly
31:39Computation amongst consumers right missing ROPs
31:42Melting connector still which really is only a 50 80 50 90 40 80 40 90 problem
31:47It's not gonna be a problem in any of these cards, especially since with the exception. I think the sapphire card
31:51They're all still standard 9 or 8 pin
31:53PCI Express cables
31:56None of the other cards from Nvidia are pulling enough power to melt the connector. So that aside though
32:00There's still a just there's still a fear of that connector being
32:03unsafe, right
32:06The
32:07The incredible tax if you will the the
32:11team green tax that kind of is carried now on the Nvidia cards, it's hard for me to recommend a
32:175070 FE or anything in the 5070 range or even 5070 TI
32:22when I know these cards are in the the price parity and cheaper and
32:27Performing just as good if not better in just our titles alone
32:30And obviously there's tons of titles out there that would that would paint a huge story
32:33So I'm curious as to honestly what some of the other reviewers are saying because I haven't talked with any of them about their
32:39Opinions on this sort of stuff, but I think the 9070 XT
32:42Specifically is a very strong competitor
32:45It's hard to recommend the 9070 though when technically it's only $50 less expensive now when you talk about features like this
32:52This is what drives the price of a card up these kinds of features
32:55Find yourself as close to SCP 6 6 or 7 599 price point that you can find on a 9070 XT
33:02And don't look back like when it comes to RT performance when it comes to upscaling when it comes to FSR improvements
33:09The gap is narrow and to be honest
33:12I
33:13Think AMD is kind of doing the right thing by sort of focusing on a smaller family of graphics cards that are as as strong
33:19As they can make them rather than trying to follow
33:21The leader when it comes to Nvidia and say how big and expensive can we make shit, but I'll be honest
33:26I do wish there was a 9080 XD because I do feel like that would be some fun
33:31Competition if you could launch that card at like
33:34749
33:35MSRP somewhere around there that could truly cause some disruption
33:39But it doesn't sound like AMD has any interest in that this time around
33:43All right guys sound off down below and this was kind of a long one
33:45There were two cards to cover but this is like the grand finale of graphics card q1 of 2025
33:52It just sort of sucks
33:53It's the first time I've actually had to talk about what a tariff and pre tariff and post tariff and worldwide
33:58MSRP versus US MSRP is and you know what that sucks big fat donkey

Recommended