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

[DaemonForce]

Who the heck got excited by the prospect of putting a 5900xt into their AM4 when the 5950x has been around for years and years???

I do. Because even if the 5900XT magically sells out after a few hours, the product launch spells out another possible price shift for similar inventory.
The 5900X and 5800X3D are still completely valid upgrade options for my daily hobbies and actual compute labor.
5600GT is probably too much compromise and the 5700X3D would be a fair third option should the rest disappear due to flash crash.
I don't have to upgrade anything but for future CPU+GPU jobs that chug, I'm better off addressing the bottlenecks now since I'm not jumping ship for AM5.
Conclusion: Super excite!

 

[fevgatos]

Huh? I just checked PCGH's Phantom Liberty test and CP2077 runs much better on a 5800x3D than any Zen 3 CPU. Like 20-30% faster. They also test worst case places.

Ram speeds. CPUs without 3d vcache get a huge penalty from ram, PCGH uses bottom of the barrel jedec officially supported ram. 4400ddr5 for alderlake, 5200 for zen 4, 3200 for zen 3. With proper ram zen 3 is within 10-15% of zen 3 x3d

I do. Because even if the 5900XT magically sells out after a few hours, the product launch spells out another possible price shift for similar inventory.
The 5900X and 5800X3D are still completely valid upgrade options for my daily hobbies and actual compute labor.
5600GT is probably too much compromise and the 5700X3D would be a fair third option should the rest disappear due to flash crash.
I don't have to upgrade anything but for future CPU+GPU jobs that chug, I'm better off addressing the bottlenecks now since I'm not jumping ship for AM5.
Conclusion: Super excite!

So the only reason you are offering is pricing. Sure, great, but you would be as excited if AMD just dropped the prices on the 5950x. That's what I'm saying, the new cpus don't offer anything other than price drops, and new cpus aren't required for those to happen.

Not everybody likes fiddling with their stuff to make it work to their liking. Not everyone is a tech enthusiast. I guess you find this hard to understand, too.

Personally, I don't mind some mild tweaking, but generally speaking, if a product doesn't work as intended out of the box, then it's not a product that's intended for my use case.

If you can't spend 3 seconds to bring a 5950x to the power level you desire and instead wait for 4 years for amd to release a CPU like the 5950x but with the power levels you want, I don't know man. I really find it absurd.

 

[AusWolf]

So the only reason you are offering is pricing. Sure, great, but you would be as excited if AMD just dropped the prices on the 5950x. That's what I'm saying, the new cpus don't offer anything other than price drops, and new cpus aren't required for those to happen.

Sure, the sole purpose of the 5900XT is to keep the 5950X's price high. You can call it scummy business tactics from AMD. I call it a win on the 5900XT's price for those looking for an upgrade. If it's literally a cheaper 5950X, then where's the harm?

If you can't spend 3 seconds to bring a 5950x to the power level you desire and instead wait for 4 years for amd to release a CPU like the 5950x but with the power levels you want, I don't know man. I really find it absurd.

Why do you think one has been waiting for 4 years? What if someone only got to the point of upgrading right now? Do you really think everyone buys everything only at product launch?

 

[fevgatos]

Sure, the sole purpose of the 5900XT is to keep the 5950X's price high. You can call it scummy business tactics from AMD. I call it a win on the 5900XT's price for those looking for an upgrade. If it's literally a cheaper 5950X, then where's the harm?

I didn't call it scummy. I'm saying it really doesn't accomplish anything by releasing the new products that they couldn't accomplish without releasing new products. At least not for the end user.

Why do you think one has been waiting for 4 years? What if someone only got to the point of upgrading right now? Do you really think everyone buys everything only at product launch?

If someone is buying right now they should completely ignore stock TDPs and buy the best they can get for their money and then spend 5 seconds to change the power limits. If the 5900xt turns out to be best for the money, buy that, if it's the 5950x, buy that. You can deal with the power limits later.

Have you ever researched what the stock brightness settings on your TV or monitor are? Never occurred to me to even bother looking for that data, since I can change the brightness.

 

[DaemonForce]

So the only reason you are offering is pricing. Sure, great, but you would be as excited if AMD just dropped the prices on the 5950x. That's what I'm saying, the new cpus don't offer anything other than price drops, and new cpus aren't required for those to happen.

Did you bother to read anything else? No? Let me break it down:
I spend a lot of time on here doing CPU jobs that barely squeak 30% total usage.
For the entire life of the system I've run an 8GB vram optioned GPU that gets majorly obsoleted in clockspeed, vram, feature level and compute by the 7900XT.
I see the writing on the wall that my overclocked 3600 is not the best CPU in the world.
In fact, look at any bench site and it will tell you "yo this CPU is bottlenecking tf out of this card!"
It's still really bad for GPU tasks like my basic ass 1080p gaming that feels best at that oddly high framerate.

You know what handles better? Literally anything with more cores.

Holy shit what an improvement. The X3D is actually worse for my situation with rates of 33-38% dropping anywhere between 16-23%.
I'm fine with the problem dropping to a window of low 20s to single digit percentages. The rest of my problems here are not fixable. ✖

 

[AusWolf]

I didn't call it scummy. I'm saying it really doesn't accomplish anything by releasing the new products that they couldn't accomplish without releasing new products. At least not for the end user.

It accomplishes having a 5950X with a different name at a reduced price.

If someone is buying right now they should completely ignore stock TDPs and buy the best they can get for their money and then spend 5 seconds to change the power limits. If the 5900xt turns out to be best for the money, buy that, if it's the 5950x, buy that. You can deal with the power limits later.

Going for the highest performing part at all costs isn't always the best idea. If a CPU is 1% slower, but 5% cheaper, I'll definitely go for that one.

Mind you, I've always been an advocate of buying non-K Intel chips because they're cheaper, eat way less power out of the box, and can be unlocked to nearly K-level performance with a power limit toggle in the BIOS. The 5900XT can also be tuned/overclocked if needed, but for an extra 100 MHz, I really wouldn't bother.

If you prioritise raw performance above all else, fair enough. I don't.

Have you ever researched what the stock brightness settings on your TV or monitor are? Never occurred to me to even bother looking for that data, since I can change the brightness.

Honestly, I don't care. I only need a TV that works and doesn't cost an arm and leg. You only see comparisons in the showroom, not in your living room.

Similarly, if the 5900XT is 1% slower than the 5950X, do you think you'll notice it in your home PC?

 

[wolf]

Holy misleading slides Batman; this is the absolutely atrocious level of misleading (potential) customers. And I keep hearing that AMD is more honest when marketing... I'll remember these slides when I hear that.

Companies aren't your friends, and AMD is speed-running the destruction of any goodwill they created.

 

[chrcoluk]

Unfortunately, until and unless AMD figures out how to put the cIOD in a low power state or even hibernation they will suffer from this.

I had a look at the TPU data and couldnt find idle power graphs, how high is AMD's typical idle power then?

The APUs are still monolithic so they're similar to Intel CPUs.

For example, on my 7950X3D, my package power at idle is around 42W total. But on my 8840U/7940HS/7840HS devices my idle can go down to a nice 2W, and 6W watching 1080p H.264 video file or a YouTube page.

The Steam Deck APU (Zen 2/RDNA2) can do the above at a locked 3W TDP APU with no stuttering. It can be a handy 6-to-8-hour media device if you're just going to watch videos.

Answers my above question, yeah that is pretty bad and shouldn't be discounted.

My PC is typically on circa 16 hours a day, and even on gaming days there may be several hours of idle or near idle loads. So that extra 35-40 watts or so will add up.

 

[lexluthermiester]

But you're not paying for a new motherboard and RAM. Hence the cost of upgradability.

Exactly! This has always been the idea behind upgradeability.

 

[AusWolf]

AMD is speed-running the destruction of any goodwill they created.

How did you come to that conclusion?

Remember that 99% of potential customers will probably never look at any marketing slides.

 

[lexluthermiester]

I honestly don't see the point in these CPUs, The 5900XT is basically a downclocked 5950X, Yippie i guess?

How do you not understand this? We're not talking about physics. We're talking about CPU upgrade model parts.

 

[AusWolf]

I had a look at the TPU data and couldnt find idle power graphs, how high is AMD's typical idle power then?

On AM5, it's typically between 15-30 W, depending on your memory configuration, IMC speed and the SoC voltage needed to run it.

 

[lexluthermiester]

AMD is speed-running the destruction of any goodwill they created.

Wait what? How do you come to THAT conclusion? Introducing more CPU choices for a platform that is at the end of it's run as a way to give people MORE value for their money is a good this. How does that hurt their consumer good will?

How did you come to that conclusion?

LOL! Ninja'd!

 

[fevgatos]

I had a look at the TPU data and couldnt find idle power graphs, how high is AMD's typical idle power then?


Answers my above question, yeah that is pretty bad and shouldn't be discounted.

My PC is typically on circa 16 hours a day, and even on gaming days there may be several hours of idle or near idle loads. So that extra 35-40 watts or so will add up.

On a single ccd chip with all background tasks off as measured from the cables a 5800x 3d draws 20w with low speed ram and 26 to 30w with high speed ram

Dual ccds with some background tasks open (hwinfo, steam etc.) can get up to 70w, tested with a 7950x. But usually you'd be around 35-40

 

[AusWolf]

Dual ccds with some background tasks open (hwinfo, steam etc.) can get up to 70w, tested with a 7950x. But usually you'd be around 35-40

There's no way you could get anywhere near that value at idle, even with Steam open. The compute die consumes peanuts at idle, having two of them can only add literally a couple of Watts max.

 

[fevgatos]

There's no way you could get anywhere near that value at idle, even with Steam open. The compute die consumes peanuts at idle, having two of them can only add literally a couple of Watts max.

Didn't say idle, that's how much my bros computer draws when booting with all background tasks. He has hwinfo, steam, nvidia broadcast, sync thing etc. Basically the same stuff I have on mine as well. I'm at 15 watts (from the normal idle that is 3-4) and his gets up to 70 from the 30w he has at actual idle.

 

[wolf]

How did you come to that conclusion?

Wait what? How do you come to THAT conclusion?

Have you seen the small print on the testing done? perhaps not as it's not shown on the slides posted here. They test CPU's with an entry level GPU (RX6600), something none of the tech press does when testing CPU's, which is what we're used to, making perhaps the entirely of the testing thoroughly GPU limited, not CPU limited.

Then the CPU's they test them against, showing their products as essentially "up to" just just as fast, sometimes faster, when TPU's own testing as well as others show that in CPU limited scenario's ZEN 3 CPU's are at serious deficits against those 13th gen Intel CPU's, like 15-20%++ depending on the game tested in CPU limited situations. Slides intended to purposely mislead consumers into thinking these XT processors are just as fast for gaming as the Intel CPU's they're tested against, when in reality there's a generational difference between them, theres no way a few hundred mhz is making up that difference.

I have no doubt the testing is real with the specs tested and literally hidden in the small print not shown here on this TPU news post, but the cherry picking and testing with an RX6600? You can't convince me this isn't a shitty thing to do and intentionally designed to make consumers think these CPU's are a gaming match for the Intel CPU's shown in the comparison.

You can argue that most potential buyers won't see these slides(which doesn't magically make it true either), but does that make the slides any less misleading to those that do see them? More CPU's for an ageing platform? cool but thoroughly unnecessary, they could just price cut existing models further, and not try to mislead anyone with the bs marketing material around new SKU's making them seem a generation better.

 

[DaemonForce]

Entry level GPUs run through games on the LOW graphical setting. I have full confidence in the numbers when setup in that way. It's nothing to worry about.

 

[wolf]

Entry level GPUs run through games on the LOW graphical setting. I have full confidence in the numbers when setup in that way. It's nothing to worry about.

Like I said, I'm sure as tested they're not outright lying, but it's still misleading.

TPU's own testing of broadly similar products bears this out, Zen3 is undeniably a generationally slower gaming CPU series than Intel 13th Gen, barring X3D models. If you're always GPU limited happy days! Sadly, that's far from always being the case.

 

[Jumbotron]

I'm still on AM4 platform. So I initially picked this news with interest. But after reading the OP, and owning a Ryzen 9 5950X, I now understand that I'm good with what I have. It would have been different if we were talking about a new 5950X3D, though.

 

[AusWolf]

Didn't say idle, that's how much my bros computer draws when booting with all background tasks. He has hwinfo, steam, nvidia broadcast, sync thing etc. Basically the same stuff I have on mine as well. I'm at 15 watts (from the normal idle that is 3-4) and his gets up to 70 from the 30w he has at actual idle.

Oh okay. Just let's not confuse that with idle, because it's not. The question was idle power consumption, not how much it eats when loading up stuff.

Have you seen the small print on the testing done?

No - I didn't even look at the slides. I know better than to give any credit to marketing crap from any company... marketing crap that, as I said, probably no one will even look at.

Marketing BS is too much beneath me to get upset about. It's laughing material at best.

 

[lexluthermiester]

Have you seen the small print on the testing done? perhaps not as it's not shown on the slides posted here. They test CPU's with an entry level GPU (RX6600), something none of the tech press does when testing CPU's, which is what we're used to, making perhaps the entirely of the testing thoroughly GPU limited, not CPU limited.

All I'm reading there is Blah, Blah and more Blah..

 

[wolf]

No - I didn't even look at the slides. Marketing BS is too much beneath me to get upset about.

All I'm reading there is Blah, Blah and more Blah..

Fantastic counter arguments to how I came to my conclusion folks, a question you both asked me. Misleading slides, confirmed.

The internet is a funny place to be sometimes

 

[lexluthermiester]

Fantastic counter arguments to how I came to my conclusion folks

Or we're tired of this same silly debate with people who refuse to see the forest for the trees. Seriously. If YOU don't see the value of what AMD is doing, that's your problem. It doesn't make US ignorant or fanboys.

 

[wolf]

Or we're tired of this same silly debate with people who refuse to see the forest for the trees.

what silly debate? you asked how I came to the conclusion, I provide my exact rationale and how I got there, you dismiss it with blah blah blah, call it a silly debate, chuck in a meaningless expression and say I'm the problem?

The slides are misleading, period. That is my only point.

Nothing you have said undoes that or convinces me that they're not, and attacking me instead of my argument isn't helping either.