[X] My Assistant

I knew epyc had one of the bigger footprints but seeing it in person is almost comedy. the thing is like 3 times the size of a regular cpu.
I wish it kept going this way and we'd have had hotplate sized ones by now, imagine the cooling surface


in other quality news one memory slot refuses to show up on my amazing bargain supermicro board even though theres no visible damage to the slot or dust, so I guess its going to be another odd numbered memory setup

throw it in an oven

Earlier this year Google nuked a $135b investment fund in Australia due to a bug in their cloud software. Lucky the Australian firm had sufficient backups, and that the bug only hit one company first allowing google to sweep through and deactivate the auto-delete timers that they accidentally left in their code base.

It seems like the new AMD CPUs aren't being so well-received, but to me the 9700X looks to be an interesting one. To match or slightly exceed the 7700X in performance (especially in productivity benchmarks, which is of interest to me) while using the same power draw as a 7700 non-X sounds fantastic, on paper at least. If only AMD wasn't so stupid with their pricing again. In my country, I can get a tray 7700 non-X for roughly half the cost of a 9700X ($299 vs $545).

"At least it's not Intel".

[www.wired.com] https://www.wired.com/story/amd-chip-sinkclose-flaw/
sounds bad, interesting to see how this will develop.

Intel microcode update is out. Mine's an Asus so gonna have to wait.

Which might be a blessing anyway, some people reported unacceptable performance loss on their chips, potentially downclocking to 800 MHz during high workload at times.

EDIT: Nvm, they just pushed it out:
- [rog.asus.com] https://rog.asus.com/motherboards/rog-maxim.../helpdesk_bios/

"Beta Version"

This post has been edited by Kagoraphobia: Aug 10 2024, 01:36

Intel microcode update is out. Mine's an Asus so gonna have to wait.

"Beta Version"

ASUS beta versions aka 'let some other people test for poison first'

*shakes fist towards the sky*

At least you can see in a 2-4 days people ranting on reddit if the new ASUS beta bios has caused any problems.


The reviews for the new AMD processors are kinda hilarious, sure intel processors are dying but the new processors are terrible because they can not beat the 7800x3D in gaming....

Just wait for the 9800x3D in a couple of months lol.

Yeah, the 7800X3D is pretty much the king when it comes to gaming. But in my opinion it's a bit of a stretch for reviewers to hammer the new CPUs for not being better at gaming than the 7800X3D when it's not supposed to be their main strength at all. In my view, the 7700/9700 are for users who want a balance between work and gaming. Gamers on a budget can go for the 7600/9600, and the 7900/7950/9900/9950 is for pure productivity.

I think the AMD marketing department screwed up royally. If it had dropped the X from the new CPUs and priced it more reasonably, I don't think we'll be seeing such scathing reviews right now.

This post has been edited by KTZ: Aug 10 2024, 03:23

Yeah, the 7800X3D is pretty much the king when it comes to gaming. But in my opinion it's a bit of a stretch for reviewers to hammer the new CPUs for not being better at gaming than the 7800X3D when it's not supposed to be their main strength at all. In my view, the 7700/9700 are for users who want a balance between work and gaming. Gamers on a budget can go for the 7600/9600, and the 7900/7950/9900/9950 is for pure productivity.

I think the AMD marketing department screwed up royally. If it had dropped the X from the new CPUs and priced it more reasonably, I don't think we'll be seeing such scathing reviews right now.

The problem is, from what I've seen the gaming performance of the 9700X is the same as the 7700X, but so is the power consumption.

It's only in MT workloads where you see a 20-40% decrease in power consumption at the same or better (especially in AVX512 workloads) performance. Which is fantastic and more of a return to the norm in my opinion, but i can see why some reviewers would call it a "flop", especially those focused on gaming. The Phoronix review (not gaming focused at all) is quite positive for example.

I've also seen some really good gains (albeit only in Cinebench sadly) when overclocking to a similar power consumption as the 7700X. Possibly AMD didn't want to "unlock" this "free" performance because raising the voltage would degrade these chips, similar to what we're seeing with Intel?

Anyway, gonna have to see what Intel does with Arrow Lake before drawing any conclusion on whether AMD fucked up with this lunch.

This post has been edited by -terry-: Aug 10 2024, 05:43

Meanwhile i'm still sitting back with my 12+ year old CPU's and laughing at everyone who's riding the upgrade train and getting burned by it, scrambling for a billion microcode updates that don't ever seem to fix anything

Also I made a website hit counter in PHP for fun the other day. Hadn't touched PHP in years or ever used it to render images before, but I was able to make it work properly in about two or three hours.
If I'd tried to do it in NodeJS I'd probably still be working on it.

But if decades' worth of AMD CPU's can really be exploited into giving you ring 0 management engine access, that could be fun to mess with. Too bad I have no AMD CPU's at all except for an 8086.

This is the thing people were warning about 20 years ago, by the way. So if it's finally happened I can just laugh even harder and hope an exploit is found for intel too so they can both drown together.

This post has been edited by Moonlight Rambler: Aug 10 2024, 08:15

Anyway, gonna have to see what Intel does with Arrow Lake before drawing any conclusion on whether AMD fucked up with this lunch.

They should just drop the fucking e-cores already. What a waste of silicon.

They should just drop the fucking e-cores already. What a waste of silicon.

[wccftech.com] https://wccftech.com/intel-bartlett-lake-s-...p-core-flavors/

I might try to deliberately boils my CPU when that generation comes out, and see Intel would offer a bartlett lake as replacement.

If they do, then the microcode is very likely not a true "fix".

My first years of using a PC was when you had to launch Windows through the DOS command prompt. Much nostalgia.

Why do my zippers keep falling down, did I lose that much weight?

They should just drop the fucking e-cores already. What a waste of silicon.

The shite e-cores are good for laptops. AMD is winning in the laptop market comparatively.

An all p-core CPU would be better for gaming performance on desktops. Intel still has not sorted the performance drop when threads are shuffled from p to e cores.

I've also seen some really good gains (albeit only in Cinebench sadly) when overclocking to a similar power consumption as the 7700X. Possibly AMD didn't want to "unlock" this "free" performance because raising the voltage would degrade these chips, similar to what we're seeing with Intel?

They are banking on 'not being intel' to move processors. They delayed launch and did a partial recall early on, it was probably to force mobo vendors to run strict stock voltages on their default settings. When people activate PBO then there is a 5-10% uplift in performance, and when you run mad unrestricted voltages on the Mobo similar to what was coming as auto then the results are similar to their marketing.

People forget for 7000 series, they had processors melting in the first month. It was resolved by micro-code and fixing the auto settings on mobos but they probably did not want to let intel off the hook by repeating the same mistake.

Its a bit of a pain that you have to wait 6months for CPU and mobo vendors to put out stable microcode before you can overclock safely, though probably not much need to overclock these days.

The launch isn't that bad but not communicated that well, they will have the normal processors for productivity / workstation and the new 3D chip for gaming.

If the new 9800X3D does not beat the 7800X3D there will be drama though. The new higher end AMD mobos will also launch around the same time as 9800X3D.

Apparently what the new Intel microcode does is to cap the max voltage to 1.55V. Nice flex tape moment there.

I manually capped mine to just 1.4V. Lost the 6 GHz boost (now "stuck" with 5.6-5.7 GHz), but honestly, I find that to be only useful for benchmark masturbation purposes. Realistically the chip ain't gonna reach that clock rate often really.

Apparently what the new Intel microcode does is to cap the max voltage to 1.55V. Nice flex tape moment there.

Some people doing in depth testing found it does some other things and can cause multicore workload instability on some mobos

Its almost like its just a stop gap measure to slow the RMAs.

I'm still of the opinion that if they forced all 13th/14th to go no higher than 1.4V, this probably would have "fixed" the issue. At least in the sense that whatever hardware design flaw(s) would have taken a long time to manifest. But gotta keep that worthless TVB performance somehow~~~~~

I'm gonna keep my voltage cap in place, until it's more or less confirmed that the RMA'd/new chips stopped having issues running at stock with the new microcode.

quite enjoying working ipmi on this new board, can get a live readout of every voltage, fan speed and temp without much oddware or even putting having to run a dedicated thing for it as on board bmc keeps track. distancing from consumer boards entirely has been enjoyable so far I think this is my one option with slightly newer technology