Sunday, June 2nd 2024

AMD Outs Ryzen 5000XT Processors for Socket AM4, an 8-year Old Socket

AMD Socket AM4 is now an 8-year-old platform, since its debut back in 2016. AMD objectively went above and beyond for this platform, launching processors powered by the original "Zen," the refreshed "Zen+," the "Zen 2," and the Intel-beating "Zen 3" microarchitecture, including 3D V-cache versions of the "Zen 3" that were competitive even with Intel's 12th Gen Core "Alder Lake" processors in gaming. Those on older processors on AM4 are spoiled for choice with upgrades within the platform, without having to change it, with AMD releasing new processor models every year for the past 8 years. The 2024 launches include the Ryzen 5000XT series.

It's hard to call the Ryzen 5000XT a "series," since there are only two SKUs—the Ryzen 9 5900XT, and the Ryzen 7 5800XT. Neither of the two feature 3D V-cache, but push clock speeds up. The Ryzen 9 5900XT is a 16-core/32-thread part, and is not meant to be confused with the 5900X, which is a 12-core/24-thread part. The 16-core 5900XT comes with a maximum boost frequency of 4.80 GHz, which is 100 MHz less than that of the 5950X. It has the same 105 W TDP, and a significantly lower $360 price. The Ryzen 7 5800XT, on the other hand, is an 8-core/16-thread chip with 4.80 GHz maximum boost frequency, compared to the 4.70 GHz of the 5800X, and the same 105 W TDP. It's priced around $260. Both chips include an AMD Wraith Prism RGB cooler that's capable of handling 140 W TDP processors. The Ryzen 9 5900XT is claimed by AMD to offer similar gaming performance to the Intel Core i7-13700K; while the 5800XT is claimed to play games competitively to the Intel Core i5-13600KF. Both chips should be available sometime in July, 2024.
Add your own comment

220 Comments on AMD Outs Ryzen 5000XT Processors for Socket AM4, an 8-year Old Socket

#201
AusWolf
wolfI have 3 AM4 boards with varying CPUs on them, one is a ryzen 1700 and could benefit from an upgrade, these slides appeared to suggest these new products are a match for 13th gen which occurred to me as an interesting proposition until I looked into it further and found it absolutely wasn't the case. So personally, AMD tried to mislead me, and for those less savvy, it's well worth making a point that the slides are misleading.

Of course you'll only see me complaining if you ignore all the others, and use just this thread as your single source of evidence, that's not really news. Maybe try other forums and social media? I already mentioned a regular person is who asked and put it on HUB'S radar to even make content about. I doubt you will because you're happy to stick to the narrative you've chosen where you think I'm the only one calling AMD out for this shitty business practise, without trying to find other ways to excuse, diminish or minimise it.
If you buy a product based on the result of one single test performed by the same company making the product, then you're an idiot. Simple as that. Don't tell me that TV adverts that claim X washing powder is three times better than the competition aren't utter bullshit. Every single company tries to catch people's attention by stupid data that doesn't make any sense. Every one of them. I don't see this situation being different.

Let's not even mention social media. People only go there if they want to show off their perfect vacations paid on finance, or if they're angry about something, which seems to be all the damn time these days.
Posted on Reply
#202
wolf
Better Than Native
AusWolfdon't see this situation being different.
I do, we're used to seeing slides from the big 3, and yet usually they're just cherry picked best case scenarios but the testing methodology itself isn't outright flawed and a significant departure from what we are used to seeing. This is.
AusWolfLet's not even mention social media
That's you're choice, but you can't deny that others are also complaining about this if you refuse to acknowledge its happening anywhere other than here, I'm not talking about Facebook BTW. Doesn't matter, you won't accept it anyway I suppose.

The lengths people will go to defend their favourite billion dollar company for free, and tbh, I'm seeing a strong correlation to the narcissists prayer here.

That didn't happen.
And if it did, it wasn't that bad.
And if it was, that's not a big deal.
And if it is, that's not my fault.
And if it was, I didn't mean it.
And if I did, you deserved it.
Posted on Reply
#203
AusWolf
wolfI do, we're used to seeing slides from the big 3, and yet usually they're just cherry picked best case scenarios but the testing methodology itself isn't outright flawed and a significant departure from what we are used to seeing. This is.
Yeah, right...






Edit: I'm trying to find marketing slides from AMD's RX 400-500 era, some of which were also utter crap.

I don't want to bring up more examples, but there's plenty from all companies. The point is: marketing is always based on flawed data and bullshit.
wolfThat's you're choice, but you can't deny that others are also complaining about this if you refuse to acknowledge its happening anywhere other than here, I'm not talking about Facebook BTW. Doesn't matter, you won't accept it anyway I suppose.
Social media is the cesspool of humanity. Whatever happens there is always taken out of proportion. And I'm not only talking about Facebook, either.
Posted on Reply
#204
wolf
Better Than Native
AusWolfYeah, right...

Yeah! right!

I'm not usually one to entertain whataboutism, but because your example is flawed when trying to use this one in particular to ...I dunno.. shoot me down?

And even if your points were absolutely bulletproof, all the more reason to blast those slides, when they drop and the news post is about them. Let them know it's not ok.




Done, up to 2x confirmed, it did the thing. Also, look at the graph carefully, yes the text says "up to 2x", but you might just notice the actual plotted points on the graph are about spot on for overall relative performance.



I remember that claim also being SHREDDED due to it not always being the case, being only so under ideal conditions, and power to those people and feeling mislead by them, call it out and ask for better. And, well at least those ideal conditions still tested a GPU against a GPU in a GPU limited test scenario. Not a CPU against a CPU in a test that doesn't test the damned CPU at all, and buries the data that sheds light on it in small print on other slides.

Excellent attempt to handwave away any other account from a source you personally won't accept too, pity that personal rule of yours doesn't bear any consequence to the facts of the matter.

What's your goal here? I walk off with my tail between my legs saying "you're right, I'm annoyed about nothing, nobody is mislead by these slides" Not going to happen, and you can just drop it anytime instead of continuing to argue a point that can't and won't be conceded, believe me it's a waste of your time.
Posted on Reply
#205
AusWolf
wolfYeah! right!

I'm not usually one to entertain whataboutism, but because your example is flawed when trying to use this one in particular to ...I dunno.. shoot me down?

And even if your points were absolutely bulletproof, all the more reason to blast those slides, when they drop and the news post is about them. Let them know it's not ok.




Done, up to 2x confirmed, it did the thing. Also, look at the graph carefully, yes the text says "up to 2x", but you might just notice the actual plotted points on the graph are about spot on for overall relative performance.



I remember that claim also being SHREDDED due to it not always being the case, being only so under ideal conditions, and power to those people and feeling mislead by them, call it out and ask for better. And, well at least those ideal conditions still tested a GPU against a GPU in a GPU limited test scenario. Not a CPU against a CPU in a test that doesn't test the damned CPU at all, and buries the data that sheds light on it in small print on other slides.

Excellent attempt to handwave away any other account from a source you personally won't accept too, pity that personal rule of yours doesn't bear any consequence to the facts of the matter.

What's your goal here? I walk off with my tail between my legs saying "you're right, I'm annoyed about nothing, nobody is mislead by these slides" Not going to happen, and you can just drop it anytime instead of continuing to argue a point that can't and won't be conceded, believe me it's a waste of your time.
It's not whatabautism. You were the one who said that we haven't seen blunders like this in marketing slides of the past. I only showed you that we actually have. Nvidia's "up to 2x" number is just as true, and just as pointless as AMD's "R7 = i7" number in a GPU-limited scenario. Neither company lied, but both took such a slim slice of the truth that made the info completely useless, and potentially misleading if you only look at marketing slides and nothing else before buying... which, again, you shouldn't do, and probably no one will ever do anyway. Tech enthusiasts know better than to trust hand-picked data, and regular Joes will never even look at the slides anyway.

Marketing is always based on bullshit. Anno 2024, this should be common knowledge. This is all I had to say. If you want to continue being angry about it, suit yourself, but leave me out of it. I've got bigger fish to fry than a few bullshit marketing slides about a product that 99% of people aren't interested in anyway.
Posted on Reply
#206
wolf
Better Than Native
AusWolfIt's not whatabautism. just as true, and just as pointles
Disagree, disagree and disagree. This blunder is far worse and markedly different. Stop trying to convince me otherwise.
AusWolfIf you want to continue being angry about it, suit yourself, but leave me out of it
You can stop replying to me anytime. I was fine with what I originally posted about it, and probably wouldn't have given it much more thought at all, until the defence force came out of the woodwork to challenge me on it. I explained, and you still can't resist trying to downplay and invalidate it ad nauseam, feel free to just.. not?
Posted on Reply
#208
AusWolf
wolfDisagree, disagree and disagree. This blunder is far worse and markedly different. Stop trying to convince me otherwise.

You can stop replying to me anytime. I was fine with what I originally posted about it, and probably wouldn't have given it much more thought at all, until the defence force came out of the woodwork to challenge me on it. I explained, and you still can't resist trying to downplay and invalidate it ad nauseam, feel free to just.. not?
Sure, you think and feel what you want, and so do I. Peace! :toast:
Posted on Reply
#209
Arcsaber
The 5800X was already daft... so they made something even dafter?
Posted on Reply
#210
lexluthermiester
ArcsaberThe 5800X was already daft... so they made something even dafter?
Wait, what? The 5800X is a wonderful CPU! What the hell are YOU on about? Making that statement you have no place calling anyone else "daft".
Posted on Reply
#211
Dr. Dro
lexluthermiesterWait, what? The 5800X is a wonderful CPU! What the hell are YOU on about? Making that statement you have no place calling anyone else "daft".
Well, I understand where they're coming from, although it wasn't a valid point when the 5800X was new. AMD withheld the launch of the 5700X until many years later (going as far as just warehousing their initial batch from 2020 to upsell the 5800X), and nowadays it's all but been eclipsed by the 5800X3D and 5700X3D CPUs. Which makes a 5800XT (basically clock boosted 5800X) at a higher price, daft indeed. It doesn't have a flagship appeal and it's effectively an old CPU for an old platform given a very thin coat of paint.

But as they say, there's no bad hardware, just bad prices :)
Posted on Reply
#212
sLowEnd
Dr. DroBut as they say, there's no bad hardware, just bad prices :)
Well, except defective and/or dangerous hardware :laugh:
Posted on Reply
#213
lexluthermiester
sLowEndWell, except defective and/or dangerous hardware :laugh:
Well there is that.
Posted on Reply
#214
lebowski89
This topic makes me laugh. Recently purchased a 5900XT, and it's working great. But I see people are doing the usual thing:

1) Assuming people are buying this to game on
2) Assuming 5950X stock is bountiful or even available (maybe it was when XT released, who knows?)
3) Assuming people are making a whole build around this CPU.

In my case, the 5900XT is for a proxmox server upgrade. I needed the extra cores. I previously had a 5600 in there. Now I know what you'll say... why not go Intel or AM5?!? My build is centred around 64GB of DDR4 unbuffered ECC ram. AMD offers the full functioning for this on non-server gear (all except MSI motherboards, pretty much). If you want ECC with Intel, you're looking at spending a mint on server gear, or you're looking at 10 year old Xeon stuff, no thanks. Any uplift going to AM5 is pretty much non-existent - I just need those cores (and DDR5 ECC is even more expensive than DDR4). As for gaming, I already have a 5800X3D and 9070XT Red Devil LE in another case - this is not a gaming chip.

Any discussions around 5900XT vs 5950x is moot in my country as of 2025 - new stock of 5950x is non-existant and the used market is also.

Inb4 Intel quick sync: I prefer Ryzen for servers. I palmed Plex off to a N100 quick sync minipc (Beelink EQ12).

TLDR: These CPU are probably not for you, but they have a use for others out there. For some, CPU like the 5900XT will tide people over for years and years for certain applications.
Posted on Reply
#215
lexluthermiester
lebowski89Recently purchased a 5900XT, and it's working great.
As it would. Just because it's one socket back doesn't mean it's useless.
lebowski891) Assuming people are buying this to game on
Why not? This CPU can do everything from 720p to 2160p without any trouble.
lebowski893) Assuming people are making a whole build around this CPU.
Again, why not? It's a great CPU for a very recent platform.
lebowski89this is not a gaming chip.
It is for many and it does very well at this. The benchmarks still prove that. One does not need an "X3D" CPU to game well.
Posted on Reply
#216
DaemonForce
The 5900X is a monster of a CPU with double the cores and threads of my older generation 3600.
It is more than double the performance at less than double the TDP. In other words, it's for work AND game.
With such an insane 12c/24t part it becomes easy to do any kind of number crunching, compiling and encode while running even a CPU intensive game.
It is a direct answer to a very power hungry problem and the 5900XT/5950X are merely insurance.
Posted on Reply
#217
lebowski89
lexluthermiesterIt is for many and it does very well at this. The benchmarks still prove that. One does not need an "X3D" CPU to game well
Indeed, and I have spent many a year gaming on non-X3D Ryzen prior to purchasing the 5800X3D - I guess what I meant: If you read the reviews and discussions about the 5900XT, people are often coming at it with a gaming workload in mind (some of this is AMD's own doing with terrible marketing that was dumped on by Hardware Unboxed and others) Sure, you can game fine on this, but you would be better served with a 5700X3D (the benefits of 3D cache are real), or even the 5800XT. No issue with those that want to game and do productivity workloads, of course. But it seems everyone focusses so much on gaming, while neglecting all the wonderful uses of those extra cores. Anyway, since purchasing the 5900XT, I've virtualised UnRaid, OPNsense, HAOS, etc, on Proxmox, and looking for more things to throw at it.
Posted on Reply
#218
Icon Charlie
DaemonForceThe 5900X is a monster of a CPU with double the cores and threads of my older generation 3600.
It is more than double the performance at less than double the TDP. In other words, it's for work AND game.
With such an insane 12c/24t part it becomes easy to do any kind of number crunching, compiling and encode while running even a CPU intensive game.
It is a direct answer to a very power hungry problem and the 5900XT/5950X are merely insurance.
This is why I have the 5900OEM. Tweaking it a bit I get 5900XT performance @65 watts. Wonderful multimedia work station.

My last computer will be built on the same concept.
Posted on Reply
#219
trsttte
lebowski89If you read the reviews and discussions about the 5900XT, people are often coming at it with a gaming workload in mind
I think the reviews and discussions are more focused on "why are they releasing this, it's just a 5900X with an extra letter at the end!?" (I know it gets slightly extra cache, which to me just indicates they're binning 5950x down to meet the 12 core demand).

It's a marketing exercise, if you wanted a 12 core cpu you could already buy 5900X for years, it probably makes sense in the frame of getting it on the headlines again and so that stores have something "new" to sell, but for people that follow tech that naturally don't care about any of that business strategy bs it just looks silly - "12core cpu, cool story but i've known about it since 2020 when you originally launched the thing".
lebowski89I've virtualised UnRaid, OPNsense, HAOS, etc, on Proxmox, and looking for more things to throw at it.
It's redundant if you already have OPNsense, but could always throw pihole in there as well, also Portainer to add which containers to launch to the which VMs to install problem :D
Posted on Reply
#220
lebowski89
trsttte(I know it gets slightly extra cache, which to me just indicates they're binning 5950x down to meet the 12 core demand).
Brother, it's a 16 core 32 thread part. Not the 12 core of the 5900X. I purchased the 5900XT because it is 16 cores and only a little more expensive than 5900X where I am (while being $100 cheaper than 5950X). And yeah, they could have just kept production of the 5950x up and dropped the prices on that, but they didn't. So I understand if people wanted AMD to go that route, but it doesn't change the fact that there are users that these CPU will be useful for. In my case, I don't really benefit from anything the 5950x offers over the 5900XT, so i opted for whatever was cheaper (while not being a used part). Also in my case, a move to either 5900XT or 5950X made sense, because I can't just go cheap out on an AM5 build. I can't just pick up whatever budget DDR5 ram or motherboard are on special. I would have to make sure that the motherboard is an X670 or X870/E, and that the DDR5 is unbuffered ECC. Just not worth it to go AM5 for me, yet. That is not to mention the increased power usage of the 7900X and co, which some people seem to care about in the server-space (not me, I got other bills I care about more than electricity).

To sum up why the 5900XT was a useful purchase for me:

- I needed a 16core/32 thread part on a platform that allows the functioning of unbuffered ECC ram on non-server parts (this rules out any notion of going to a 13700k or whatever)
- I needed a part for a productivity-focused build, I already have a 5800X3D for gaming (aka if I needed a 5700X3D I would have got one)
- Any uplift of 5950X or AM5 means little for my intended workload - my upgrade to AM5 or AM6 for this build will require $$$ due to my server/unbuffered ECC requirements.
- I don't care about Intel or any potential reduction in idle power usage of Intel - and I certainly don't like that big and little core rubbish they do - my server was always going to be AMD.

I'm more frustrated in those thinking everything is about gaming. I can understand people questioning why these parts exist in the first place.
trsttteIt's redundant if you already have OPNsense, but could always throw pihole in there as well, also Portainer to add which containers to launch to the which VMs to install problem :D
Yeah, I have Debian VM with all my docker containers (including Portainer, Traefik, Authelia, etc, etc), in which I use Ansible + Docker Swarm to set it all up. Been meaning to set up pihole :)
Posted on Reply
Add your own comment
Jul 12th, 2025 00:13 CDT change timezone

New Forum Posts

Popular Reviews

TPU on YouTube

Controversial News Posts