[a / b / c / d / e / f / g / gif / h / hr / k / m / o / p / r / s / t / u / v / vg / vr / w / wg] [i / ic] [r9k / s4s / vip / qa] [cm / hm / lgbt / y] [3 / aco / adv / an / asp / bant / biz / cgl / ck / co / diy / fa / fit / gd / hc / his / int / jp / lit / mlp / mu / n / news / out / po / pol / qst / sci / soc / sp / tg / toy / trv / tv / vp / wsg / wsr / x] [Settings] [Search] [Home]
Board
Settings Home
/g/ - Technology


Thread archived.
You cannot reply anymore.



File: 10_side.jpg (42 KB, 1024x768)
42 KB
42 KB JPG
>his RAM isn't directly mounted on his processor
>>
All of /g/ is firmly mounted on your mom
>>
>>69764584
>Can't upgrade RAM and CPU seperately
Oh /g/ee tell me more anon
>>
>>69764584
The Raspberry Pi Zero does this but with BGA packages
>>
>>69764659
Are you retarded?
The brown thing between the two chips is a socket.
>>
>>69764584
Can you do away with RAM altogether and use the CPU cache only?
>>
>>69766101
There's 3 ways of doing that:
>Write a tiny ass program that fits in the cache
>add fuckloads of cache, which is bloody expensive
>effectively replace the cache with RAM, removing any benefits
>>
>>69766339
>Write a tiny ass program that fits in the cache
Most programs should be like this.
>>
>>69764584
SoC
>>
>>69766101
If one type of storage was fast enough, you could do a computer on a chip. Imagine 1GB of CPU cache used as RAM, storage, and cache. ~200gbps speeds.
>>
just put some HBM on the processor dude
>>
>>69764697
this
>>69764584
in the pic the socket is supposed to be for a rom anon
>>
File: 1546756508723.gif (822 KB, 600x366)
822 KB
822 KB GIF
good thread
>>
>>69766339
>add fuckloads of cache, which is bloody expensive
The more cache you add, the slower it gets.
>>
>>69767913
i still dont know why this is not happening.
amd has all the assets to make it possible plus it would eliminate all latency issues.
>>
>>69764651
fpbp
>>
>>69764584
Literally Faggin processor
>>
>>69764584
HM6116P-3
2048-word X 8bit High Speed CMOS Static RAM
>>
>>69764584
It is.
All phones do that.
>>
>>69768806
Because resulting chip would be bigger, thus more difficult to manufacture and more expensive.
According to rumors nvidia is also looking into a more modular design.
>>
>>69770430
Uh uh so akin tel me why you think Intel doesn't just stick a fuck ton of level 3 cache on its processor where the fixing igpu is. The igpu takes up the vast majority of the die
>>
>>69770397
No.
>>
>>69766101
certain microprocessors use a scratchpad or a tcm, which is basically cache-speed RAM but directly addressed rather than transparently mirroring main memory
>>
>>69766339
Something like Ebin could store an entire OS in its cache.
>>
>he doesn't wire his own cpu using ttl chips
>pleb
>>
File: file.png (302 KB, 1154x391)
302 KB
302 KB PNG
>>69764584
But anon, it is.
>>
>>69764651
Bloooooooowwwwwwnnnn up



Delete Post: [File Only] Style:
[Disable Mobile View / Use Desktop Site]

[Enable Mobile View / Use Mobile Site]

All trademarks and copyrights on this page are owned by their respective parties. Images uploaded are the responsibility of the Poster. Comments are owned by the Poster.