Jump to content

IPC = instructions per clock. GHz is an indicator of the clock. They are two different measures. Multiply the two and you have an indicator of the throughput. IPC isn't usually given a value though, and it will vary with task and with a lot of other factors so is hard to nail down. You might see generalisations comparing architectures.

Gaming system: R7 7800X3D, Asus ROG Strix B650E-F Gaming Wifi, Thermalright Phantom Spirit 120 SE ARGB, Corsair Vengeance 2x 32GB 6000C30, MSI Ventus 3x OC RTX 5070 Ti, MSI MPG A850G, Fractal Design North, Samsung 990 Pro 2TB, Alienware AW3225QF (32" 240 Hz OLED)
Productivity system: i9-7980XE, Asus X299 TUF mark 2, Noctua D15, 64GB ram (mixed), RTX 4070 FE, NZXT E850, GameMax Abyss, Samsung 980 Pro 2TB, iiyama ProLite XU2793QSU-B6 (27" 1440p 100 Hz)
Gaming laptop: Lenovo Legion 5, 5800H, RTX 3070, Kingston DDR4 3200C22 2x16GB 2Rx8, Kingston Fury Renegade 1TB + Crucial P1 1TB SSD, 165 Hz IPS 1080p G-Sync Compatible

Link to comment
https://linustechtips.com/topic/781885-ipc-vs-speed/#findComment-9857674
Share on other sites

Link to post
Share on other sites

IPC is directly related to the processor architecture itself . Generally , adding more hardware increases IPC . My disdain for the term aside , it's defined as the number of operations the processor can perform in one cycles , ie one hertz .

Clock speed refers to the number of cycles the processor will go through per second . Please note that clock speed does NOT refer to the speed the transistors  switch at .

One big problem is that power consumption and heat do not scale linearly with clocks . It grows exponentially. That's a meain reason why we see clock speeds topping out a 4-5ghz .

Generally , with high IPC , you cannot run at such high clocks because the extra hardware generates more heat .

It's a compromise the designers make when building the architecture.

It's a serial approach vs a parallel approach ( to some extent )

AMD Ryzen R7 1700 (3.8ghz) w/ NH-D14, EVGA RTX 2080 XC (stock), 4*4GB DDR4 3000MT/s RAM, Gigabyte AB350-Gaming-3 MB, CX750M PSU, 1.5TB SDD + 7TB HDD, Phanteks enthoo pro case

Link to comment
https://linustechtips.com/topic/781885-ipc-vs-speed/#findComment-9857707
Share on other sites

Link to post
Share on other sites

2 minutes ago, Coaxialgamer said:

One big problem is that power consumption and heat do not scale linearly with clocks . It grows exponentially. That's a meain reason why we see clock speeds topping out a 4-5ghz .

Power consumption does scale with clock... if you keep the voltage constant. It is the increasing voltage required with clock that drives the power non linear.

Gaming system: R7 7800X3D, Asus ROG Strix B650E-F Gaming Wifi, Thermalright Phantom Spirit 120 SE ARGB, Corsair Vengeance 2x 32GB 6000C30, MSI Ventus 3x OC RTX 5070 Ti, MSI MPG A850G, Fractal Design North, Samsung 990 Pro 2TB, Alienware AW3225QF (32" 240 Hz OLED)
Productivity system: i9-7980XE, Asus X299 TUF mark 2, Noctua D15, 64GB ram (mixed), RTX 4070 FE, NZXT E850, GameMax Abyss, Samsung 980 Pro 2TB, iiyama ProLite XU2793QSU-B6 (27" 1440p 100 Hz)
Gaming laptop: Lenovo Legion 5, 5800H, RTX 3070, Kingston DDR4 3200C22 2x16GB 2Rx8, Kingston Fury Renegade 1TB + Crucial P1 1TB SSD, 165 Hz IPS 1080p G-Sync Compatible

Link to comment
https://linustechtips.com/topic/781885-ipc-vs-speed/#findComment-9857840
Share on other sites

Link to post
Share on other sites

Example: i5 vs Celeron D, both running at 3.0Ghz

 

They have the same clock speed, but the i5 has like 10 times IPC more than the celeron D.

 

Therefore, the i5 will give you more performance than the celeron D, while running at the same clock speed.

 

The difference is in the architecture

CPU: Ryzen 2 2700@ 4.0Ghz    Mobo: Gigabyte X470 Gaming 7 Wifi    Cooler: EVGA CLC 240    GPU: GTX1080 FTW DT @ 2113Mhz   PSU: EVGA 750W P2   

Case: Fractal Design Meshify C   Displays: 34" LG34UC79G, 24" Dell

Link to comment
https://linustechtips.com/topic/781885-ipc-vs-speed/#findComment-9857891
Share on other sites

Link to post
Share on other sites

42 minutes ago, porina said:

Power consumption does scale with clock... if you keep the voltage constant. It is the increasing voltage required with clock that drives the power non linear.

Yes. I know. But realistically that doesn't happen. 

AMD Ryzen R7 1700 (3.8ghz) w/ NH-D14, EVGA RTX 2080 XC (stock), 4*4GB DDR4 3000MT/s RAM, Gigabyte AB350-Gaming-3 MB, CX750M PSU, 1.5TB SDD + 7TB HDD, Phanteks enthoo pro case

Link to comment
https://linustechtips.com/topic/781885-ipc-vs-speed/#findComment-9858064
Share on other sites

Link to post
Share on other sites

IPC is Instructions Per Clock. For every cycle, the chip will execute a certain number of instructions. And bear in mind that a cycle is nothing more than an operation that a chip is performing over and over again. 

 

If you have a chip that operates at a speed of 4GHz(or 4 billion operations per second), and an IPC of 50, you multiply 4,000,000,000 by 50, and you have the total number of instructions the chip could execute per second. However, if the code is optimized for a particular architecture(for example, if you make it smaller), you could effectively increase the IPC, delivering better performance.

"It pays to keep an open mind, but not so open your brain falls out." - Carl Sagan.

"I can explain it to you, but I can't understand it for you" - Edward I. Koch

"I didn't die! I performed a tactical reset!" - Apollolol

Link to comment
https://linustechtips.com/topic/781885-ipc-vs-speed/#findComment-9858081
Share on other sites

Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

×