Why no more odd numbers (5, 7 or 9) core CPUs? AMD had a tri-core.

>>58983834
because cores are on the die in pairs, so you can't disable a single core, you have to disable them in pairs.
Also, making straight tri-core CPU's doesn't make sense, because it wastes a bunch of die space. Binary systems work more efficiently in multiples of 2.

Because hexagonal CPU dies are kinda a stupid idea

Weren't AMD's Tri-Cores just quad-cores that had one of the cores fail in production?

>>58983869
Yeah, same way that 6-cores are just 8/10/12 core CPU's that have one of the core pairs defected (or just disabled, because, you know, profit margins).

>>58983869
Disabled yes, fail maybe. I have a tricore that has been running all 4 for years.

>>58983869
Yes.
The failures were usually minor (Semi conductor producers are EXTREMELY strict when it comes to failures) so you could reactivate the disabled core and have it work perfectly fine most of the time.

>>58983869
Yes, it was exactly that, and they just disabled the dead one and rolled it out as a tricore.

One would presume they decided to stop sucking at silicon fabrication since then.

>>58983944
Nah, those types of issues happen no matter how good you are at silicon.
Silicon fabbing is a fucking crapshoot. They'll zap an entire silicon wafer with lithography stuff and make like 400 dies on it, and then cut it up and test them and only like 20 will work properly and pass all the tests.
It's getting worse as we get to smaller process nodes.

>>58983851
Then why make tri core in the first place?

>>58983951
Nobody "makes" a tri core die.
They're all 4 core dies.

>>58983951
Try reading.
They didn't.

>>58983964
But he said you can't disable a single core? That means it had to have been made with the intention of being a tricore?

>>58983974
The cores on Zen are slightly different so they need to be disabled in pairs, but traditionally disabling one core is what goes.

>>58983974
He was talking about modern cores. That's why there are no 7-Cores which are defective 8-cores, only 6-Cores.

Back then the cores weren't linked the same way and you could disable a single one.

>>58983884
Are you retarded?
6 is an even number and can be perfectly placed on a chip.
5 or 7 would be be the weird thing.

Who /unlocked/ their tricore phenoms?

File: 1480948523974.png (1MB, 1856x2592px) Image search: [Google]

1MB, 1856x2592px

>>58983834
Because 3 != shitpost

>>58983964
Xbox 360 CPU was a tricore, no fourth core anywhere

>>58983834
>AMD had a tri-core.

because 1 core was dead on the chip

File: 1484879314247.png (118KB, 1218x743px) Image search: [Google]

118KB, 1218x743px

WHATS GOOD NIGGAS

had a 4870 originally but got a 7950 a few months back for free

File: xbox360_processor_die.jpg (51KB, 379x450px) Image search: [Google]

51KB, 379x450px

>>58984361
Weird. Well, it's possible, but not normal.

>>58984381
>3 core cpu
>3gb gpu ram
>6gb ram
>300gb disk

>>58984381
i like your (multiples of) 3 autism

keep it up

>>58983974

The first multicore cpu's from amd were composed of individual cores. Tri cores were nothing more than quad cores with a "defective" one, that was disabled.
Quoted, because more often than not amd couldn't keep up with demand, so they started using perfectly good cpu's, with cores disabled. This caused the whole unlocking craze back then, more notoriously dual cores unlocked to quad ones without issues.

File: 1422069055157.jpg (67KB, 363x1082px) Image search: [Google]

67KB, 363x1082px

>>58983834
>AMD had a tri-core.

>>58984145
I'm using one right now, since I sold my 6700K this week to prepare for Ryzen. An X3 720 that I bought in 2009. Fourth core works fine. It was never a good overclocker though, so it's only at 3.2GHz. Can't touch the northbridge frequency at all, even with voltage. Not the motherboard either, as I had a 1090T running it on it at 3.9GHz/2800MHz/1.425V for a time.

>>58984370
Not necessarily. Many (if not most) were functional, with some more stable than others. There were also batches where perfectly functional cores were disabled to meet demand for the lower part. There was a certain dual core Phenom II that unlocked to a perfectly-functional full quad core >90% of the time. Plus the later Thuban-based quad cores that unlocked to hex cores.

Man, I loved the Phenom II range. Shame it was never quite there in terms of raw performance.

>>58985726
How is that trolling? If you have three functional cores there's literally no reason not to sell it as a tri-core.

>>58985749
That image was made when that face was called coolface and not trollface.
It's a usage of that face from before you reddit tier shitlords can even comprehend.

>>58985794
I'm so terribly sorry I haven't been using the same website as long as you have
>reddit
Never been there, don't associate me with that cancerous shithole.

>>58985730
>Shame it was never quite there in terms of raw performance
It ended up being faster than the core 2 arch
At least the cpus that didnt have shitload of cache

File: >she.jpg (78KB, 1280x720px) Image search: [Google]

78KB, 1280x720px

>>58985794
>reddit tier shitlords
Your attempted deception has not gone unnoticed.

>>58986097
The problem being that Core 2 was launched in mid-2006 and Phenom II in early 2009, several months after the first-gen i7s.

>>58983834
amd had tri-cores because somethimes in process of production one core was turning out to be worse. you can unlock many x3's and get four cores, but usually fourth core is barely keeping stable on stock clocks, not to mention overclocking.

>>58984112
It can be, but very rarely is. On pretty much every modern CPU, it's just a 8 core part with 2 disabled.