General CPU Questions

[Fez760]

Now I have an understanding and the basics about a CPU but i still don't know a lot of it and im pretty sure knowledge would help my Questions are..

 

Why more than one core?

 

What is the significance on the number of threads

 

operating and max turbo frequency? -is this just like speed?

 

L2 and L3 Cache?????

 

64-Bit Support? or is it exactly how it sounds

 

Support For Virtual Technology? what on earth does that mean!

 

Graphics Base and max dynamic frequencies? i literally feel like that's just speed...

 

Integrated memory controller speed? Does this like help regulate Ram?

 

Thermal Design power???

[Guest acdcman200]

more cores help your pc do more than one thing

turbo frequency is how fast your cpu will clock up to if its not running over temp

64 bit support is for running 64bit applications and supporting more ram

thermal design is how the cpu handels heat like tdp

 

Now I have an understanding and the basics about a CPU but i still don't know a lot of it and im pretty sure knowledge would help my Questions are..

 

Why more than one core?

 

What is the significance on the number of threads

 

operating and max turbo frequency? -is this just like speed?

 

L2 and L3 Cache?????

 

64-Bit Support? or is it exactly how it sounds

 

Support For Virtual Technology? what on earth does that mean!

 

Graphics Base and max dynamic frequencies? i literally feel like that's just speed...

 

Integrated memory controller speed? Does this like help regulate Ram?

 

Thermal Design power???

[minibois]

-snip-

- Because programs can use more than one

- Threads are like virtual cores kinda. Every core = 1 thread. If you have a CPU with hyperthreading, every core = 2 threads.

Of course, threads are good because programs can use more than 1 (or more than that)

- Operating frequency is the speed (typically measured in GHz) of a CPU. Turbo boost is something on Intel only IIRC, which makes your CPU's clockspeed (frequency) higher if necessary.

- Not sure how to explain the rest correctly

[NeatSquidYT]

more cores help your pc do more than one thing

turbo frequency is how fast your cpu will clock up to if its not running over temp

64 bit support is for running 64bit applications and supporting more ram

thermal design is how the cpu handels heat like tdp

1. Kinda

2. Doesn't really make sense

3. Yeah

4. Thermal design power IS TDP, and it's not really how it handles heat.

[Fez760]

- Because programs can use more than one

- Threads are like virtual cores kinda. Every core = 1 thread. If you have a CPU with hyperthreading, every core = 2 threads.

Of course, threads are good because programs can use more than 1 (or more than that)

- Operating frequency is the speed (typically measured in GHz) of a CPU. Turbo boost is something on Intel only IIRC, which makes your CPU's clockspeed (frequency) higher if necessary.

- Not sure how to explain the rest correctly

But you did explain what you could well but uhm IIRC?

[JamGorby]

But you did explain what you could well but uhm IIRC?

 

TL;DR incoming. Please hold........

[minibois]

But you did explain what you could well but uhm IIRC?

IIRC = if I remember correctly.

 

And this video may help:

[Glenwing]

Now I have an understanding and the basics about a CPU but i still don't know a lot of it and im pretty sure knowledge would help my Questions are..

 

Why more than one core?

 

What is the significance on the number of threads

 

operating and max turbo frequency? -is this just like speed?

 

L2 and L3 Cache?????

 

64-Bit Support? or is it exactly how it sounds

 

Support For Virtual Technology? what on earth does that mean!

 

Graphics Base and max dynamic frequencies? i literally feel like that's just speed...

 

Integrated memory controller speed? Does this like help regulate Ram?

 

Thermal Design power???

Multiple cores are important to a point, they help the computer process multiple things simultaneously, which is important for modern software. It's not the only thing that matters though, and having more cores will only help you up to a certain amount. After that you are making sacrifices in power and clock frequency and not really getting much in return.

Every core can process one thread at a time. Assigning additional threads helps increase CPU efficiency.

Frequency is not a measure of speed. Like I said in your GPU thread, the CPU is regulated by an internal clock. Every time the clock ticks, each core on the CPU does whatever it does once. How much work it gets done during each tick depends on the design of the CPU. High frequency alone doesn't mean the performance will be good, and for that matter neither does the number of cores. The core design plays a large role in performance, and it's not something you can quantify on a spec sheet, at least not in a readable way. For this reason, as with GPUs you need to compare CPUs with benchmarks, not specs.

Cache is a small section of the CPU for storage of data it's currently working on. It's used in the same way as RAM, but it's much faster but also much smaller capacity. There are several levels of cache in modern CPUs, the L1 cache is part of the core now, each core also gets a dedicated L2 cache, which is larger but slower than the L1 cache. Depending on CPU design all cores may share an L3 cache which is larger but slower than the L2 cache, and if you have an integrated GPU you may even have an L4 cache which is shared between the CPU and integrated GPU, which is larger but slower than the L3 cache. Beyond that, you have the RAM, which is much larger but much, much slower than the cache.

64-bit support is part of any modern CPU. Basically it means it supports modern software.

VT-d is some random enterprise-grade feature, don't worry about it.

The memory controller is a part of the CPU that controls communication between the CPU cores and RAM. It can handle up to a certain frequency by default, but can usually go beyond its rated spec if you want.

TDP is the max power consumption under sustained heavy load, at default settings.

[JamGorby]

snip

The first two questions kind of go hand in hand. On a single processing core, multithreading is generally implemented by time-division. That means the CPU is switching to one thread, doing some work there, and then switching to the next, and etc etc. based on the threads priority. Priority determines how long that thread can occupy the CPU. With a only handful of apps open (or threads to process) this can happen so fast it creates the illusion that the CPU is doing multiple things at once. In reality, a single core can still only handle one thread at a time. What happens if the OS needs to deal with hundreds of threads? The amount of time the CPU spends with each thread before coming back to, say, the thread that's letting you type a "b" in your browser's search bar will become longer to the point where there's noticeable lag. Adding more cores is like adding more CPU's where there used to be only one (this is simplified of course, as there's hyperthreading and AMD's modular design, etc.) Now the OS can divide these threads up among the cores. Even though each of these cores can still only do one thing at a time, they can work in tandem so when you type "b" in your browser the amount of time it takes for the CPU to get to it is less and you're not sitting there wondering if you need to hit that key again. (we've all been there). 

Turbo Boost allows the CPU to dynamically raise it's clock speed when the CPU isn't under heavy loads. Say you have a 4790k and you are doing something that's loading up one core heavily. Turbo Boost will allow that core to raise it's clock speed to 4.4 to do the work faster provided thermal and power requirements are met. If you have a load on 2 cores, it can run those cores at up to 4.4 (but usually 4.3) if temps and power are still good. If you load the entire CPU it will stay at it's base clock of 4.0 to ensure stability, save power and produce less heat. It's sort of a way for the CPU to intelligently overclock itself on the fly based on workloads. 

Caches are low capacity bits of memory built into the die of a CPU. They are much smaller than system memory but multitudes faster. Really small bits of repetitive code that will be frequently accessed while doing a job can be stored here so the CPU can fetch it much quicker than it can from system memory to increase performance. L1/L2/L3 generally go from smaller and faster to bigger and slower respectively. Also, depending on the way the CPU is made, You may have a small fast L1 cache dedicated to each core with 2 larger but slightly slower shared L2 caches for each pair of cores and a large l3 cache accessible buy all cores that's even bigger yet slower. 

Think of it in terms of having a work crew loading tools from a shop (ram) into a utility truck (L3) to go do some work. Team one and team two each grabs a toolbox from the truck (L2) and go to work while wearing their own individual tool belts (L1). 

It's faster to grab a wrench from your tool belt than to borrow one from your buddy. But if you don't have one, its faster to borrow his than it is to go across the street to see if you can borrow the other teams tools. But if no one has a wrench they'll have to go to the truck to get one which will take more time. But, it's still faster than going back to the shop to grab one. 

The main benefit of 64 bit CPU's is to allow access of more than 4GB of ram. Even still, on modern 64bit CPU's with 64bit OS's, 32bit programs, while they will run, only have access to 4GB of ram. A 64bit application can access more. Much more. How much more? 16 exabytes (theoretically). What's an exabyte? It's ...um.....a lot. 

Support For Virtual Technology means that the CPU has a special instruction set to help it run virtual machines better (VT-x on Intel, AMD-v for...well...you know). Virtual machines let you run....virtual machines. Ever wanted to split your quad core 16gb ram computer into 4 1 core 4GB machines and have them be treated like separate machines so you can run linux on your linux while your linux is running linux? That's what VM's do. 

Graphics Base and max dynamic frequencies = Basically Turbo Boost for video cards. 

Memory controllers used to live in the chipset. The part of the chipset that contained the memory controller (and some other things) was/is called the Northbridge. When you see one of us geeks talking about NB speed/frequency/voltage/etc. we're talking about the Northbridge. The Northbirdge now resides on the CPU itself for both Intel and AMD. AMD was the first to put the NB on the chip and use an IMC (integrated memory controller). Intel followed suit later. Having the memory controller on the chip itself allows for much faster access to RAM than having a separate chipset for this function. The trade off is they sacrificed some flexibility. Back in the old days, you could upgrade to a new RAM standard by upgrading the chipset i.e. motherbaord. Now, to support new memory standards, the CPU and motherboard must both support that standard. Hence the reason you can't use DDR4 on AM3+/LGA1150 boards. Even if the board was made to physically support it, the CPU's do not. Now that we've grown accustomed to things, it was a worthy trade off. 

TDP is the maximum amount of heat generated by the CPU that the cooling system in a computer is required to dissipate in typical operation (thanks wiki). While it can be used to loosely judge how much power a CPU can use, it is not the amount of power a CPU will use as is so commonly misconstrued.

Basically, if you designed a CPU that can use 200watts at load and you find you must be able to absorb 100watts of that energy into something (heat sink)  to keep your CPU at acceptable operating temps, you have a 100watt TDP CPU. (I think...lulz).