0.8ghz or 8ghz?

[RTX 3070 I7 11800h]

Hey, I know this isnt real but I was curious. Would a 80 core 0.8ghz (0.01ghz a core) Be more powerful then a single core 8ghz processer (I am curious lol)

[minibois]

1 minute ago, RTX 2070 Max-Q said:

Hey, I know this isnt real but I was curious. Would a 80 core 0.8ghz (0.01ghz a core) Be more powerful then a single core 8ghz processer (I am curious lol)

There is no one answer.

Let's assume a world where these two products can actually exist and let's assume the IPC (instruction per clock, 'what it can do per Hz') is the same:

 

it will still depend on the program.

Is the program heavily parallelized or not?

 

I think most programs would run faster on a 8Ghz 1 core CPU, just because it can do the tasks one after another. But some tasks don't require a single task to go one after another, but may need multiple tasks done at the same time.

[NicoV]

Probably depends on the workload. For the 1 core you run into the problem of scaling with frequency, the higher the speed of a single core, the harder it is to feed that core with information. For the 80 core, you don't need to give each core as much data, making the pipeline much easier, but it is hard to make something that scales well with 80 cores.

[RTX 3070 I7 11800h]

1 minute ago, NicoV said:

Probably depends on the workload. For the 1 core you run into the problem of scaling with frequency, the higher the speed of a single core, the harder it is to feed that core with information. For the 80 core, you don't need to give each core as much data, making the pipeline much easier, but it is hard to make something that scales well with 80 cores.

 

3 minutes ago, minibois said:

There is no one answer.

Let's assume a world where these two products can actually exist and let's assume the IPC (instruction per clock, 'what it can do per Hz') is the same:

 

it will still depend on the program.

Is the program heavily parallelized or not?

 

I think most programs would run faster on a 8Ghz 1 core CPU, just because it can do the tasks one after another. But some tasks don't require a single task to go one after another, but may need multiple tasks done at the same time.

I mean like gaming cinebench r20 and stuff like that

[Coolmaster]

Look into Amdahl's law for how parallelization and speed interact.

[RTX 3070 I7 11800h]

1 minute ago, Coolmaster said:

Amdahl's law makes it so that parallelization will be slower than sequential of the same speed.

Same speed?

[minibois]

2 minutes ago, RTX 2070 Max-Q said:

gaming

probably single 8Ghz core

2 minutes ago, RTX 2070 Max-Q said:

cinebench r20

Single core test: 8Ghz CPU

Multi-core test: equal.

[RTX 3070 I7 11800h]

1 minute ago, minibois said:

probably single 8Ghz core

Single core test: 8Ghz CPU

Multi-core test: equal.

Decent. 6 core 5ghz i7 10750h with a RTX 2070 Max-Q

 

EDIT: is enough for me lol

[Vishera]

 

9 minutes ago, minibois said:

There is no one answer.

Let's assume a world where these two products can actually exist and let's assume the IPC (instruction per clock, 'what it can do per Hz') is the same:

 

it will still depend on the program.

Is the program heavily parallelized or not?

 

I think most programs would run faster on a 8Ghz 1 core CPU, just because it can do the tasks one after another. But some tasks don't require a single task to go one after another, but may need multiple tasks done at the same time.

Well,it's possible to answer it if you know the related equations.

Quote

IPC = X

8GHz = 8000MHz
0.8GHz = 800MHz

1 MHz = 1 million Hertz

 

80 cores 0.8GHz:
800,000,000X*80 = 64,000,000,000X

 

64,000,000,000X Instructions per second

 

Single core 8GHz
8,000,000,000X Instructions per second

 

[Hairless Monkey Boy]

14 minutes ago, RTX 2070 Max-Q said:

Hey, I know this isnt real but I was curious. Would a 80 core 0.8ghz (0.01ghz a core) Be more powerful then a single core 8ghz processer (I am curious lol)

Well, if you had done your math correctly, and both CPUs had the same number of total clock cycles per second, one on a single thread, the other split between 80 threads, then the single core CPU would almost always be faster, because there is an overhead penalty for multi-threading.

[minibois]

12 minutes ago, RTX 2070 Max-Q said:

Decent. 6 core 5ghz i7 10750h with a RTX 2070 Max-Q

 

EDIT: is enough for me lol

Wait, I answered your question assuming it was the same cumulative Ghz.

Now I see this:

25 minutes ago, RTX 2070 Max-Q said:

80 core 0.8ghz (0.01ghz a core) Be more powerful then a single core 8ghz processer

The single core 8Ghz would be faster than an 80-core 0.8Ghz CPU.

I assumed your question would be 

1 x 8(Ghz) = 8Ghz

vs.

80 x 0.1Ghz = 8Ghz

 

[RTX 3070 I7 11800h]

3 minutes ago, minibois said:

 

I assumed your question would be 

1 x 8(Ghz) = 8Ghz

vs.

80 x 0.1Ghz = 8Ghz

 

No I was curious

[-rascal-]

42 minutes ago, RTX 2070 Max-Q said:

Hey, I know this isnt real but I was curious. Would a 80 core 0.8ghz (0.01ghz a core) Be more powerful then a single core 8ghz processer (I am curious lol)

 

The frequency (GHz) doesn't get divided into the number of cores.

When you say 0.8 GHz, it's 0.8 GHz for all cores.

 

Ryzen 5600X is 6-core 12-threads, and with a 3.7 GHz Base Clock.

All 6-cores are running at 3.7 GHz .... not 0.6166 GHz per core, for a total of 3.7 GHz.

 

@Vishera's answer would be correct.

 

[Mark Kaine]

That's basically what Sony and IBM tried with the PS3. 

 

The pioneers of computing! 

[Mark Kaine]

35 minutes ago, HairlessMonkeyBoy said:

Well, if you had done your math correctly, and both CPUs had the same number of total clock cycles per second, one on a single thread, the other split between 80 threads, then the single core CPU would almost always be faster, because there is an overhead penalty for multi-threading.

But... multi threading can do things at the same time... I always thought it'd be faster, more difficult to program for (but not impossible obviously) and, faster (because it can do multiple things at the same time, also if you only have 1 core you'll get issues with background processes quickly, I'd guess!) 

[Vishera]

6 hours ago, Mark Kaine said:

That's basically what Sony and IBM tried with the PS3. 

 

The pioneers of computing! 

Not really,the CELL BE had 8 SPEs and 1 PPEs.

That CPU was designed to process both graphics and CPU calculations at the same time,

So in addition to the 7800 GTX based GPU in the PS3 you can use the CPU for graphics as well.

6 hours ago, Mark Kaine said:

But... multi threading can do things at the same time... I always thought it'd be faster, more difficult to program for (but not impossible obviously)

The days of hard coding a program to use a specific amount of cores are over,

Now days things are programmed to dynamically scale with the amount of cores you have.

Blender is a good example for it.

[Hairless Monkey Boy]

2 hours ago, Mark Kaine said:

But... multi threading can do things at the same time... I always thought it'd be faster, more difficult to program for (but not impossible obviously) and, faster (because it can do multiple things at the same time, also if you only have 1 core you'll get issues with background processes quickly, I'd guess!) 

Here's some data for you.

 

Methodology:

Restricted number of active cores in BIOS, restricted Frequency with Windows power plan. Verified the correct frequencies and number of cores with HWinfo.

 

To start, I restarted the PC, then closed all background programs including Afterburner, ICUE, AISuite, Malwarebytes, etc. I opened Cinebench R23, and used task manager to set it to high priority. Then I let it run for 10 minutes under both conditions and recorded the final score.

 

System:

See specs for BabyBlu in my sig. All settings were held constant except for core count and core clock.

 

Results:

8 cores and 16 threads running at 2.6Ghz:

Score: 7167

 

4 cores and 8 threads running at 5.2Ghz:

Score: 7090

 

Conclusion:

You can see that even though the two configs have theoretically the same total number of clock cycles per second, that the configuration with fewer total threads ekes out a lead of 77 points or 1.09%.

[Guest]

3 hours ago, RTX 2070 Max-Q said:

Hey, I know this isnt real but I was curious. Would a 80 core 0.8ghz (0.01ghz a core) Be more powerful then a single core 8ghz processer (I am curious lol)

I've read this thread straight 3 times. 

 

0.01ghz a core? total 0.8ghz total on 80 cores? 

 

80 cores 0.8ghz total = 800mhz

so 0.01 is 10mhz.... per core.

 

You're talking about 80 calculators on a single die there.

 

The single core at 8 ghz would smoke in in all productivity. Even multi tasking.

 

 

[Vishera]

4 hours ago, ShrimpBrime said:

0.01ghz a core? total 0.8ghz total on 80 cores? 

That is indeed an unusual way to word it.

4 hours ago, ShrimpBrime said:

You're talking about 80 calculators on a single die there.

 

The single core at 8 ghz would smoke in in all productivity. Even multi tasking.

You are correct:

Quote

 

IPC = X

8GHz = 8000MHz
0.01GHz = 10MHz

1 MHz = 1 million Hertz

 

80 cores 0.1GHz:
10,000,000X*80 = 800,000,000X

 

800,000,000X Instructions per second

 

Single core 8GHz
8,000,000,000X Instructions per second

 

The single core 8GHz processor,assuming the IPC is the same,is exactly 10 times more powerful than the 80 core 10MHz CPU.

[Guest]

1 hour ago, Vishera said:

That is indeed an unusual way to word it.

You are correct:

The single core 8GHz processor,assuming the IPC is the same,is exactly 10 times more powerful than the 80 core 10MHz CPU.

Would probably take like 800 minutes before you saw a glimpse of desktop lol. lmao. omg. imagine that.

 

And I thought a 4 hour WPrime 1024m took a long time on a PIII Coppermine 256 at 450mhz 

 

[Vishera]

9 hours ago, ShrimpBrime said:

Would probably take like 800 minutes before you saw a glimpse of desktop lol. lmao. omg. imagine that.

 

And I thought a 4 hour WPrime 1024m took a long time on a PIII Coppermine 256 at 450mhz 

 

A Pentium 3

I wonder how far you can push it with an overclock,after all you are a HWBot user

Is that Windows XP?

I have been trying to run XP on my Ryzen machine but can't make it work.

[Guest]

31 minutes ago, Vishera said:

A Pentium 3

I wonder how far you can push it with an overclock,after all you are a HWBot user

Is that Windows XP?

I have been trying to run XP on my Ryzen machine but can't make it work.

That was an OEM build from a thrift store some years back for like 5 bucks. = no OC.

Mobile nanoBGA.

 

Yep XP. And running socket 939 on the bench currently on XP.

 

For your Ryzen build, you want to use windows 7. Turn on USB legacy compatibility so you can use the keyboard and mouse.

 

Find yourself USB drivers that work in previous gen board for W7, should work ok.thats what I did on my Maximus X hero for benching 3D.

[Vishera]

11 minutes ago, ShrimpBrime said:

For your Ryzen build, you want to use windows 7. Turn on USB legacy compatibility so you can use the keyboard and mouse.

 

Find yourself USB drivers that work in previous gen board for W7, should work ok.thats what I did on my Maximus X hero for benching 3D.

I already did that,just without the USB legacy compatibility,the chipset drivers fixed it for me.

But your solution is superior.

[Guest]

40 minutes ago, Vishera said:

I already did that,just without the USB legacy compatibility,the chipset drivers fixed it for me.

But your solution is superior.

Hey as long as it works....

 

XP is so dated, it's really only useful these days for legacy boards. I cant even get a working web browser lol.

[Vishera]

45 minutes ago, ShrimpBrime said:

XP is so dated, it's really only useful these days for legacy boards. I cant even get a working web browser lol.

Well,i like to experiment with different operating systems,

and i have seen that Windows XP is really good for RAM benchmarks,

Top results are on Windows XP,the current world record:

https://hwbot.org/submission/4446283_safedisk_maxxmem_ddr4_sdram_5375.5_marks