Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)

[DisEnchantment]

Speculate at will

 

[Thibsie]

5700X3D is cheaper and only 5 to 10 fps slower. If you REALLY need an X3D. But if you get it, you might be able to keep your system till AM6 arrives

That may be true. I always figured the loss of frequence going from 5600X to 5700X3D was a bit of a let down.

 

[igor_kavinski]

That may be true. I always figured the loss of frequence going from 5600X to 5700X3D was a bit of a let down.

But you gain two extra cores/four threads.

 

[Schmide]

It's hard to make a full comparison. I spent most of the time setting it up before going back. They are both well provisioned. The laptop (Lenovo LEGION 5 Pro 15.6" Gaming Laptop - AMD Ryzen 7 7735HS - GeForce RTX 4060 - 165Hz WQHD (2160 x 1350) - Windows 11 - Storm Gray) has 64gb ram a 1tb and 4tb drive. The 5800x3d system has many drives ~20tb, a 3070ti and a 360 aio. It doesn't ramp up like the laptop and there are always enough ports for everything. There was a small time where I had the laptop running my monitors and it did a decent job. The only small issue is it only has one display port and hdmi 2.0 so the display is limited. This kept me from running HDR and full res on multiple monitors at the same time. Also I like to run hdmi out to the tv on a 50ft fiber optic cable and that had to be swapped out every time. One weird thing is the laptop seems to constantly send a wake signal to the tv and the tv asks if I want to switch to the input even if it isn't on. Desktop doesn't do this.

I could probably get by on the laptop but my daily driver system is comfortable.

I promise I will move to the 9800x3d system someday.

Edit: oh yea. The 5800x3d system is windows 10 and windows 11 sucks. I couldn't watch the superbowl live yesterday so I taped it. I decided to play with the 9800x3d system before going to bed and windows 11 decided to spoil the game by showing the score on the load in screen. I disabled that after but it was already too late. (insert bad language forum exception here) microsoft.

 

[Tuna-Fish]

At Computex AMD displayed a slide about plateforms longevity, nothing about AM4 but given how it still sell it could extend a little after the 2026 stated commitment, for AM5 they displayed 2027+, wich mean that they dont really know when it will be EOL and that it could be extended to 2029-2030 depending of the sales.

Not just sales, memory standard transitions can be tricky, with delays fairly late in the game. This is why both manufacturers have in the past built CPUs that support two different memory standards; it's not just a way of increasing consumer choice, it's also about getting the new stuff when it's hot and also risk-reducing in case the new stuff is delayed.

 

[eek2121]

Puget Systems Most Reliable Hardware of 2024

With 2024 wrapped up, we want to take the opportunity to call out specific brands and models that had exceptional reliability over the last year.

www.pugetsystems.com

SHOCK. SURPRISE.

Is AMD sprinkling gold dust on these CPUs?

Personally, I think that the fact that each TR CPU having a fuse blown when it is overclocked, is keeping their users from pushing them beyond comfortable levels.

They have fuses? That is both dumb and doesn’t really change anything. Here in the U.S. at least, they would still have to cover any warranty claim unless they can show that overclocking caused the damage.

I meant that the noise is a problem when using laptop since my desktop is so quiet under full load that it doesn't even matter that I use open-back headphones. My laptop is really noisy when doing almost nothing. Just a single gif in a Slack message makes the fan go crazy.

So, I have hard time believing that there's any reasonable sized desktop replacement laptops that doesn't sound like a vacuum cleaner.

Sounds like you have a poor quality laptop.

Well a power supply for a desktop replacement gaming laptop is (I have one) from around half Kg to a Kg for 280/330W type, the most modern ones can go on the lower side. GaN is rising even there. Same power adapter was around one Kg just a few years ago. USB-C is OK up to 100W, there are ways to push it to 240W and that involves pushing the voltage to 50V and special cables but definitely for the highest performing notebooks it is not enough, as they need even more than that. And a 240 W USB-C PD supply will not be light.

A lot of it comes down to the design of the laptop and the charger. There are 250+ W chargers out there that are much smaller and lighter than the one that comes with laptops.

I was just looking on Amazon and I found a laptop charger that weighs about a pound, supports 330+W, includes a bonus USB charging port (charge your phone AND you laptop!) and is quite small, 1/3rd the size of most power bricks.

There are laptops that draw more than 240W?

Do they even make 240w USB-C PD supplies with a single output? All the ones I've seen are large because they have multiple ports - maybe that's because they're so big anyway they might as well use the extra real estate, so I don't know how small they could be made. Or if there is enough potential customers for a single port model designed for minimum size to be worth marketing.

IIRC USB-PD 240W is a newish standard and there aren’t many options.

Not just sales, memory standard transitions can be tricky, with delays fairly late in the game. This is why both manufacturers have in the past built CPUs that support two different memory standards; it's not just a way of increasing consumer choice, it's also about getting the new stuff when it's hot and also risk-reducing in case the new stuff is delayed.

AMD’s biggest mistake, IMO, was not supporting DDR4.

 

[igor_kavinski]

AMD’s biggest mistake, IMO, was not supporting DDR4.

Didn't cost them that much. AM4 is still selling. I can understand why they did it. Less validation and support headaches down the road. If they had supported DDR4 on AM5, they would have needed to halt AM4 CPU production immediately and move to expensive AM5 production which would have reduced their profits.

 

[coercitiv]

at least 30 systems

This sounds like a commitment issue. Have you experienced a catastrophic system failure as a kid?

 

[leoneazzurro]

A lot of it comes down to the design of the laptop and the charger. There are 250+ W chargers out there that are much smaller and lighter than the one that comes with laptops.

I was just looking on Amazon and I found a laptop charger that weighs about a pound, supports 330+W, includes a bonus USB charging port (charge your phone AND you laptop!) and is quite small, 1/3rd the size of most power bricks.

Yes, you can go full GaN and get a more compact adapters, but most producers keep saving on the costs there: even if the cost of these materials is getting down the difference with silicon is far from being negligible. Also, beware that power supplies designed for gaming laptops are designed to handle also transients with peaks above the rated value, while other commercial power supplies usually indicate the maximum combined peak value.

 

[DAPUNISHER]

Cleaned up the the last couple of pages after getting a number of reports. Please stay on topic, and stop making it personal.

Mod DAPUNISHER

 

[eek2121]

AMD needs to drop the 9950X3D soon. I need to give them more money! 🤣

 

[Tup3x]

Sounds like you have a poor quality laptop.

ThinkPad T14 Gen 1 (Renoir based). If it wasn't work laptop, I would have at least repasted it but it has always been hot and noisy. Weird considering the low TDP. I could get replacement if I just asked since it's not exactly new anymore. Still waiting for better options to show up.

 

[igor_kavinski]

ThinkPad T14 Gen 1 (Renoir based).

4800H? That's pretty old. I almost bought one a few years ago but the nagging thought of it being Zen 2 prevented me from going ahead. Also, 6000 series was the one that got AMD's special power saving sauce so get it replaced with a 6900HX or 6800HS one. I've used a 5825U one and I could barely hear the fans on that one. It was a 13.3 inch HP Aero.

 

[igor_kavinski]

Not a good comparison (only 9950X3D score at the moment) but at least it is showing some uplift for the X3D, assuming both 5080's are stock: https://browser.geekbench.com/ai/v1/compare/198431?baseline=194858

 

[igor_kavinski]

This very bad: https://videocardz.com/newz/redditor-reports-amd-ryzen-7-9800x3d-burned-while-watching-tv-series

 

[Tuna-Fish]

This time there doesn't seem to be any damage in the socket plastic bits, so it was presumably installed correctly. I wonder if there was a bent pin, some conductive trash forming a bridge between the pins, or if there was just a manufacturing error in the CPU or the socket itself.

 

[igor_kavinski]

some conductive trash forming a bridge between the pins

We can rule that out I think. Wouldn't have lasted 20 days unless there's some process that causes unconductive crap to turn into conductive crap through heat over a period of three weeks. Most likely ASROCK BIOS defaults or EXPO voltage handling are to blame.

 

[Tup3x]

4800H? That's pretty old. I almost bought one a few years ago but the nagging thought of it being Zen 2 prevented me from going ahead. Also, 6000 series was the one that got AMD's special power saving sauce so get it replaced with a 6900HX or 6800HS one. I've used a 5825U one and I could barely hear the fans on that one. It was a 13.3 inch HP Aero.

Close, Ryzen 7 PRO 4750U. Starting to feel sluggish. Strix Halo would be nice.

 

[lightmanek]

We can rule that out I think. Wouldn't have lasted 20 days unless there's some process that causes unconductive crap to turn into conductive crap through heat over a period of three weeks. Most likely ASROCK BIOS defaults or EXPO voltage handling are to blame.

There is also a possibility of manufacturing defect in the motherboard, socket or CPU itself.

Electronics is quite sensitive to bad contact. A bit of finger grease on the pad can cause resistance to raise dramatically, high resistance will cause higher heat and from there we can have cascade failure.

 

[lightmanek]

Close, Ryzen 7 PRO 4750U. Starting to feel sluggish. Strix Halo would be nice.

I rock 4700U on a HP laptop and I love it (as a package). Have newish 6800U laptop for over a year now but somehow I still prefer my trusty HP ENVY 15 as it performs well enough for my work tasks.

 

[fastandfurious6]

This time there doesn't seem to be any damage in the socket plastic bits, so it was presumably installed correctly. I wonder if there was a bent pin, some conductive trash forming a bridge between the pins, or if there was just a manufacturing error in the CPU or the socket itself.

what are the chances the burn is falsified? is it easy to do?

It's strange how the 'burned' cpus and gpus since last year are by far the most popular models

nvidia GPUs sure have a real problem as there's massive juice going into them,

but 9800x3d? not 9900x, not 9950x... and was it 7800x3d that burned last year? can't remember which one


and realistically.... "while watching tv series" wtf. allegedly not overclocked and barely 10% utilization lol. somehow 9800x3d got so thirsty for watts overnight, bypassed all failsafes, went all the way and burned and crashed itself lol. it's just so unrealistic

 

[Tuna-Fish]

what are the chances the burn is falsified? is it easy to do?

It would be, but there would be little point.

These kinds of failures have always been happening, at a steady low rate. CPUs are mass-produced products, they have non-zero failure rate.

The difference is that now they get more press, because there were some much more widespread failures, and now every news item about it gets clicks.

 

[igor_kavinski]

and realistically.... "while watching tv series" wtf.

Must've been a heck of a scene. Too much for the 9800X3D to handle. I wonder if it was Trump telling a joke in The Apprentice

 

[fastandfurious6]

It would be, but there would be little point.

These kinds of failures have always been happening, at a steady low rate. CPUs are mass-produced products, they have non-zero failure rate.

The difference is that now they get more press, because there were some much more widespread failures, and now every news item about it gets clicks.

yeah but realistically how many cpus running stock non-OC near-idle were reported 'burned' out of nowhere on the internet during let's say 2010-2022? do you remember any example?

definitely saw super-OCed get burned but stock and 10% ? 🤔

 

[igor_kavinski]

From the dead 9800X3D reddit thread:

 

[fastandfurious6]

smells like FUD