TSMC dont actually play that game. Why, bcoz their leading the node shrink pack so without throwing prices they ady got a line of customers willing to throw money for each new node.

What they do is preferential block batching, where a the first batch nodes will be sold at the highest markups (invariably Apple came with the first pick), and then the markup goes down as subsequent batches are fulfilled. Others playing catchup will charge less but its the release of these new nodes which is important as many tied their product launches to their CPU SOC which is based on each cutting edge node iteration. No fool will launch a flagship phone using CPU based on older nodes, good luck if they do coz its simply bad marketing.

Look back at your own postings to know. You made your opinion, okay, so just leave it or else your just trolling here

Stock setting is 253W TDP, mobos auto OC and push the CPUS to surpass that. If you manually set TDP to its rated max, no issues. So it that CPU fault or mobo fault?

And NO, 7800X3D exploding CPU was not just due to OC, redditor Skyfishjy had this to say:



So NO, AMD case wasnt better. It was just as bad as Intel today

Anything that is made must have a set max limit, CPUS, even super hypercars. Why makers have set electronic limited 300kmph, or 400kmph, is not bcoz the car engine or aerodynamic cannot push it further but bcoz other things ie tyres, gearbox, etc, are not as durable to reach such high speeds. But what happens when the makers offer you an unlimited Ksku hypercar that could go as fast as you can floor the pedal? Well obviously something else happen if you try to drive like that on a daily basis, rite?

The same logic goes with Intel Ksku CPUS, the speed limits are off, the brakes are gone if you dont set any restriction how fast you want to push. If you set your own PL1=PL2=max PBP limit, many dont seem to have this CPU degradation problem.

Is it an Intel fault for unlimited Ksku version CPU? Yes
Is it mobo maker fault for unlimited power setting? Yes
Is it the user fault for not limiting their CPU? Yes
Is it market fault for creating OC'able CPU version? Yes
Is it Intel fault for not telling buyers to get nonK if they dont know what their doing? Yes

When everyone point fingers are each party, everyone is actually at fault too. Peeps should just buy nonK and leave the Kskus to wizards like derbauer or Kingpin. They arent complaining about hardware failures when pushed beyond the limits.

This post has been edited by babylon52281: Jul 17 2024, 11:48 AM

X86 also has a lot of legacy baggage in tow. Imagine that a 14900K could technically run programs made for Intel 8008 CPU, thats crazy longtime to be keeping compatibility. Many software have already ditched 32bit so in order for X86 to move forward at the least Lunarlake should ditch 32bit ISA and purely run on 64bit with software emulation to handle for any remainder 32bit programs.

Yes, Intel tried that with Itanium but then the emulation layer was terribly overcomplex while the CPU then werent fast enough for seamless software translation. These days, CPU are so efficient and powerful compared to IA64 that any emulation for less taxing 32bit shouldnt be a problem anymore.

On a hardware level, legacy eats up transistors which could be put to use for more modern usage rather than wasting away. Directly it is bigger than needs to be, indirectly it might be why X86 has the poorer power efficiency as it needs to keep these legacy hardware "warm" in case they are called.

Indeed Win11 makes everything worse, too bad we need to change by next year anyhow while Win12 still not here yet (and could be worse than 11!). As for Edge its not surprising as its based off the same Chromium now. Use Mozilla or Opera instead.

If want hybrid low power use, Intel has Alder-Raptors design which the Ecores does perform efficiently but the problem is the BIG Pcore which loses in IPC vs AM5/6. And then rather than increasing it remains as 8 Pcores while Intel keep spamming Ecores which is POS when calls for heavy multitasked workloads vs AMD R9s. Imho just to run Windows & other background tasks, it only needs 4-8 Ecores so Raptor should have been configured as 10 Pcores + 8 Ecores giving the R9s a good fight.

SD so far been underwhelming, I think its coz Win11 is not designed for such low power SOC. Rather than Windows it needs a native ARM based OS.

Its still a HX CPU, so its not such a good idea to let it boost to 157W in such a small chassis.

CPU degradation due to excess heat & voltage will affect all to certain levels but on laptops with its poorer heat transfer the effects could happen earlier. This is provided if HX CPU behaves like Ksku where it will ramp above max PBP baseline. As with desktop Ksku my advice would be to manually turn off auto OC and limit PL1 & PL2 to its official max TDP.

Yes and no. The 'fix' will prevent degradation issues on fresh builds or those to be installed (make sure to update bios) as long as you ALSO MUST disable mobo auto OC as it will still break thru the limits set by Intel anyways. Here is what I mean (see timestamp 17:16)


but for those ady with problems, once its broken there is no repair, just RMA it.

13/14Gen nonK dont have these issues...

And AMD also have their own problems

7000series no major issues? Dont believe all the spiel that fanboys tells you


And 9000series delay due to bugs caught the very last minute up to post shipment. What others still not yet found? As you said we dont really know...

There are no "safer" routes. Want safe & proven, go 12Gen or AM4.

I heard some hardware (recent Xeons iinm) needs a special softkey to unlock full potential so it may be that certain hardware needs persistent internet connection to have full access? But that would be really scummy similar to many AAA games.

Companies like to milk as much money from customers. If they could do that to server market, whos to say they wont try it with consumer market?
Game industry is prime example, what used to be fluff addons DLCS in some games can be the difference between winning & losing. While it may not be true for now, its uncertain what might be for near future CPUS or even GPUS (possible future might sell you a RTX 9900 for the price of 9800 but you have to pay yearly subscription to unlock to 9900)

It took them months to fix last I heard. And it affects all X3Ds but of course 7800X3D are most affected due to best seller. Well if one can forgive AMD certainly same too for Intel, rite?

Intel, like AMD, are large corporations with risk of lawsuits if its not fixed permanently. Lets wait and see once the microcode is roll out, meanwhile no need to wait for the fix to arrive, just pull back everything like I said above and it will help save it sooner.

Tell that to Intel. IQ no meed much?

I will reply to both with the same answers. Basically if I were to trust the internetz then AMD would have been finished with their own SOC overvoltage burn issue as well the conspiracies from 9000series delayed launch, as well I would be with Intel and this 13/14 Gen issue.

Architectually, 13/14 RPL isnt that much different that 12 ADL, so rather if you could get a good deal for 12700/12900 then why not, for the price beats later gen i5 CPUS. So get it more for the price performance rather than any concern.

If your a regular user with no intention to OC, get the regular nonK 13/14Gen. The PL2 drop to 65W is there protect. Even if you unlock PL2 limit, just remember to disable mobo auto OC, and set PL1 & PL2 to Intel PBP limit.

If your set on OCing and you realised its why you get a Ksku then you be knowing what your doing & how to minimise the degradation effects. So its a risk you knowingly take yeah.

Intel extended warranty means squat if users are lazy to do those above & mobos continue to push those CPUS beyond limit.

As for options, AM4 is an option but it lags in power efficiency to LGA1700, and still continues to bug by more unstable platform. AM5 is still overpriced due to mobo & DDR5.

This post has been edited by babylon52281: Aug 4 2024, 01:20 AM

Err a windows update crash has very little to do with the CPU, so blame should be given but given to the right causes. Its the same like peeps blaming Microsoft for Crowdstrike outtage, and while the its easy to blame Windows for being too sensitive to kernel faults, the blame should be going to CS for for their buggy update.

My 2sen experience, SSD seems very sensitive to hard crashes I had a MLC drive that supposed to be more durable but it died after few Windows crash. If your game is this sensitive, I suggest run it from a HDD as it has less damages from crashes, my HDDs that was connected to that MLC drive outlasted and still in use today.

From what I read, X3D has poorer idle power and has higher power usage in light use, understandable as its a gamer CPU rather than the universal CPU that is Intel monolithic. My 12700F does ~3W in internet use, X3D i read does 3X more. Gaming per watt does beat Intels which is where Ryzen does better. If usage doesnt benefit the cache, its no better or worse than the nonX versions. Its good in its niche role tho.

Decent AM5 mobos still not as cheap as decent LGA1700 mobos tho and Intel still can leverage cheaper but speedy DDR4. DDR5 doesnt have a speed over poor latency advantage until it hits 6000MHZ which is still pricier than DDR4. And if your getting Aseries mobo it will really hamper any CPU upgradability due to poorer VRMs, I would not reco pairing with current or future X3D for example, a good starter for upgrade purpose is to go for a B650 mobo with decent VRM. Much like Intel Hmobos, AMD Amobos are meant for office use with little to no intention for CPU upgrade.

But if one has the money for a decent build, yeah go for AM5. Intel Arrowlake will be new evolution so will have issues to resolve, wait for Lunarlake at least once their chiplet tech has matured.

As for preference, I would rather Intel to put the locks on their mobo partners eagerness to push CPU beyond limits (im looking at Asus, MSI as the worst). Also Intel should have hard limits even on Ksku that doesnt allow it to be driven beyond.