Best processor for max performance

[Gorgon]

No, you're just not understanding.

No, YOU are not understanding. You can't just add the speed of CPU's or CPU cores together and say well this thing is a bajillion Ghz. That's bullshit.

Bandwidth is depending on many other things, most importantly the bus to the memory. That's bandwidth. You're just full of shit putting people on the wrong foot with your misinformation.

[selig]

No, you're just not understanding.

No, YOU are not understanding. You can't just add the speed of CPU's or CPU cores together and say well this thing is a bajillion Ghz. That's bullshit.

Bandwidth is depending on many other things, most importantly the bus to the memory. That's bandwidth. You're just full of shit putting people on the wrong foot with your misinformation.

Allow me to edit your post to reflect the tone we strive for here at ReasonTalk:

"You can't just add the speed of CPU's or CPU cores together and say well this thing is a bajillion Ghz. That's incorrect.

Bandwidth is depending on many other things, most importantly the bus to the memory. That's bandwidth."

I hope you can see the difference, and agree that in both cases you have stated your case, but in your "version" you will likely provoke a "like" response (read: negative name-calling distraction to the topic), while in my version you simply set the record straight and move on.

So if your intention is to be provocative and evoke a "like" negative response, your comments are not welcome here. But if your intention is to correct the record and present factual evidence to back up your claim, you will need to learn to edit yourself. Your choice, and I KNOW you are aware of what comes next if you choose the former over the latter.

That is all.



Sent from my iPad using Tapatalk

[Gorgon]

Point taken. Thanks Selig.

[XysteR]

No, you're just not understanding.

No, YOU are not understanding. You can't just add the speed of CPU's or CPU cores together and say well this thing is a bajillion Ghz. That's bullshit.

Bandwidth is depending on many other things, most importantly the bus to the memory. That's bandwidth. You're just full of shit putting people on the wrong foot with your misinformation.

I've already explained as much as I'm prepared to, to help you understand. It's not misinformation - I've explained everything if you even bothered to read. It has to be said, I'm not taking your attitude towards me (or others in this forum) lightly either. I've noticed this in your other posts here, and to be honest it stinks! You're the exact type of negativity that existed in the props forum, before the likes of your type turned it on a downward spiral and the reason why I never bothered to go there. It just takes one bad apple I suppose.

Selig gave you an option here and you've blatantly ignored his advice.

[selig]

No, you're just not understanding.

No, YOU are not understanding. You can't just add the speed of CPU's or CPU cores together and say well this thing is a bajillion Ghz. That's bullshit.

Bandwidth is depending on many other things, most importantly the bus to the memory. That's bandwidth. You're just full of shit putting people on the wrong foot with your misinformation.

I've already explained as much as I'm prepared to, to help you understand. It's not misinformation - I've explained everything if you even bothered to read. It has to be said, I'm not taking your attitude towards me (or others in this forum) lightly either. I've noticed this in your other posts here, and to be honest it stinks! You're the exact type of negativity that existed in the props forum, before the likes of your type turned it on a downward spiral and the reason why I never bothered to go there. It just takes one bad apple I suppose.

Selig gave you an option here and you've blatantly ignored his advice.

Not sure if you saw his last post, I gave him an option and and he replied "Point Taken, Thanks Selig (which I TOTALLY appreciate, BTW).

Should I ALSO edit YOUR post to demonstrate a better way of communicating the point you're trying to make? Or were you intentionally being ironic?

With that in mind, no, it does not HAVE to be said, at least not in this thread. Instead I would prefer you contact a mod if you have a complaint, otherwise this thread will more likely go off the rails in my experience.

Thanks for listening, and for (hopefully) understanding.

[HepCat]

Whilst l did not foresee this argument, l'd like to add that l've always thought of multicore CPU speed rating in a similar way as parallel resistors.

OK it's very little like a parallel resistors equation, except l picture the cores to be in parallel and l envisage the CPU speed to be no greater than the speed of the fastest core.

Here's how l imagine CPU cores:

CPU speed = Hz = frequency = a rate
Thus a CPU can be like a mill wheel, which as far as the river running through it is concerned, is determined by its rate.

Let's say we have 5 waterwheels spanning the breadth of a river. Each water mill spins at 2 Hz.

The total MILL "speed" for the entire arrangement is thus 2 Hz, because nothing is spinning faster than that; in fact at (2 x 5) = 10 Hz, any wheel would splinter into pieces.

However, because we have 5 wheels in the arrangment, it's extremely powerful, compared to 1 wheel spinning at 2 Hz.

Another analogy: 5 arms flexing at 2 Hz each, = 2 Hz cycle speed, but with the innate strength of 5 arms.


This is actually special to me because l used to sell PCs and l got irked when rival sellers would rate their Core 2 Duos at 6 gHz when l, and the manufacturer, would rate it as Core2Duo 3 gHz. Once l got a call from a prospective customer and was asked to explain the difference in CPU freqs (i.e. which one was correct and why) and l couldn't say, and despite the other seller's 6 gHz offering blatantly being a 3 gHz Core2Duo, l was, strangely, unable to assert that we were both selling more or less the same machine (because l couldn't reason my conclusion properly).

[Gorgon]

Here's how l imagine CPU cores:

This is where it goes wrong. I'm not a CPU designer or CPU scientist but I have years and years of experience with building computers and closely monitoring performance. Comparing a CPU to a windmill is very primitive, to say the least.

Let me take my example, where I said that if you have 8 cars in parallel going 50 miles per hour, is not the same as going 400 miles per hour. Now, if you took the 8 engines, and build some very ingenious gearbox where all that engine power comes together, and you also build a chassis around it that can hold that power and can also transfer that particular power to the wheels, you might be able to reach what, maybe 300 miles per hour.

But that's not how CPU's or CPU cores work. They are engines, they do not have a common gearbox, or crankshaft. They are lonesome cores of power that are depending on smart programming to take advantage of all the combined CPU power.

There are no programs that are so cleverly programmed that they can do this. From what I understand, any particular Reason device chain runs on one core. Which means that one core has to wait until the other core is done with whatever it is they're doing. That is your equivalent of those 8 cars going in parallel. Sure, the core SPEED makes a difference in when a particular task is done, but you cannot add the cores together to come to somekind of total sum which is larger than the elements. It's just nonsense.

Hope I explained that right.

[HepCat]

Here's how l imagine CPU cores:

This is where it goes wrong.

How is my scenario dissimilar to the one you're pleading?

[Gorgon]

How is my scenario dissimilar to the one you're pleading?

I think I said so, comparing it to a windmill. Mechanical comparisons to electronical comparisons. I'm not exactly sure but your post wasn't exactly inviting to analyze because it just doesn't make any sense on a basic physical level.

[HepCat]

How is my scenario dissimilar to the one you're pleading?

I think I said so, comparing it to a windmill. Mechanical comparisons to electronical comparisons. I'm not exactly sure but your post wasn't exactly inviting to analyze because it just doesn't make any sense on a basic physical level.

I explained that both deal with rates (frequencies). We are trying to rate a CPU, by assigning it a frequency. That's the bottom line isn't it.

By the way, in both cases, something flows through. In the case of the 5 waterwheels, you have water flowing through in virtual lanes or threads (correct term? I mean, the east side of the river has little impact on a wheel on the west side of the river). In the case of the CPU, you have information flowing through the cores.

I should have said that the analogy is with a mill wheel, not a mill. The mill = the CPU, the waterwheel = a CPU core.


By the way, there was also the analogy of 5 arms flexing at 2 Hz per arm. The result is a flexing frequency of 2 Hz, but all the power of 5 arms.

[Gorgon]

I think we're deviating from the point. The rule of thumb here (boondock saints reference) is that more is better.

Better CPU score? More performance. CPU marks are a reliable benchmark to measure raw CPU performance. Obviously the whole computer comes into play (motherboard/memory/chipset) but the CPU marks give you a number on how the raw performance of the CPU is. That is a clever algorithm which gives you performance in a number of different tasks, such as floating point operations (through the whole CPU).

So like I said, more is better. The whole performance of the computer depends on a lot of other factors, such as RAM speed, BUS speed, having an SSD or not, and the brand and type of the chipset.

[XysteR]

Whilst l did not foresee this argument, l'd like to add that l've always thought of multicore CPU speed rating in a similar way as parallel resistors.

OK it's very little like a parallel resistors equation, except l picture the cores to be in parallel and l envisage the CPU speed to be no greater than the speed of the fastest core.

Here's how l imagine CPU cores:

CPU speed = Hz = frequency = a rate
Thus a CPU can be like a mill wheel, which as far as the river running through it is concerned, is determined by its rate.

Let's say we have 5 waterwheels spanning the breadth of a river. Each water mill spins at 2 Hz.

The total MILL "speed" for the entire arrangement is thus 2 Hz, because nothing is spinning faster than that; in fact at (2 x 5) = 10 Hz, any wheel would splinter into pieces.

However, because we have 5 wheels in the arrangment, it's extremely powerful, compared to 1 wheel spinning at 2 Hz.

Another analogy: 5 arms flexing at 2 Hz each, = 2 Hz cycle speed, but with the innate strength of 5 arms.


This is actually special to me because l used to sell PCs and l got irked when rival sellers would rate their Core 2 Duos at 6 gHz when l, and the manufacturer, would rate it as Core2Duo 3 gHz. Once l got a call from a prospective customer and was asked to explain the difference in CPU freqs (i.e. which one was correct and why) and l couldn't say, and despite the other seller's 6 gHz offering blatantly being a 3 gHz Core2Duo, l was, strangely, unable to assert that we were both selling more or less the same machine (because l couldn't reason my conclusion properly).

Yes this was a bit of a marketing thing back in the day particularly of the dual core CPUs - I was in the same boat as you. We couldn't get to 6Ghz because of thermal issues, so it was decided to run multiple cores at a thermally manageable frequency.

I like your analogy - If we look at torque to be the main power at the spindle end of your wheel then we get 10 Hz because the spindle can handle that amount of torque, that's the designed strength of the spindle. 1 waterwheel attached to the spindle powered by the river would indeed not have as much torque as 5 waterwheels attached to the spindle - This would be an ideal though to get exactly 5X torque. There is a clutch at the end of the spindle, the clutch being other system bottlenecks. But the clutchs' working design is clever, in that the clutch tries to give us as much of the 5X torque as it possibly can. Hehe I think this is going a bit too far.

[XysteR]

How is my scenario dissimilar to the one you're pleading?

I think I said so, comparing it to a windmill. Mechanical comparisons to electronical comparisons. I'm not exactly sure but your post wasn't exactly inviting to analyze because it just doesn't make any sense on a basic physical level.

It makes perfect sense, semiconductor transistors are indeed quantum mechanical - His analogy is valid.

[Gorgon]

I like your analogy - If we look at torque

Uh yeah.

[XysteR]

I like your analogy - If we look at torque

Uh yeah.

Torque being my simile of total CPU power if you didn't understand

[Gorgon]

Torque being my simile of total CPU power if you didn't understand

Nope. Because it doesn't make any sense at all. If torque would be something that you could compare with a computer part, it would probably be something in a conventional harddrive.

[XysteR]

Torque being my simile of total CPU power if you didn't understand

Nope. Because it doesn't make any sense at all. If torque would be something that you could compare with a computer part, it would probably be something in a conventional harddrive.

It makes perfect sense - you're just not seeing it. An analogy or simile is not meant to work the way you think. This may come as a shock, but get this. We're allowed to use anything we choose! Wooo lol.

Anyway, like I said earlier. This is completely away from the OP now.

[Gorgon]

Anyway, like I said earlier. This is completely away from the OP now.

You mean, like *I* said earlier.

[XysteR]

Anyway, like I said earlier. This is completely away from the OP now.

You mean, like *I* said earlier.

No like I said earlier, page one three posts from the bottom, end of post.

[gak]

I'm not sure I understand this thread, but please, for the sake of reasonable argument.......understand.

I'm not some shcmuck who hasn't been through all this. It's not an exact science, but there are some important things. Reason doesn't deal with CPU use the way every other host does (nor does "every other host" )

My system currently: 4770k, not overclocked, MSI board (much better than the gigabyte I recently had , seriously) and performance is..........good. But considering the careful information here, I think that there are more powers at be than the OP specified.

Still, regardless of anything, fwiw, my personal opinion is that GHZ is king, and overclocking is the second best denominator. Sorry. But I don't see any evidence to the contrary.

YMMV. I really do hope that you get the best possible scenario. Just speaking from experience.

[Gorgon]

Still, regardless of anything, fwiw, my personal opinion is that GHZ is king, and overclocking is the second best denominator.

There were 3Ghz CPU's 10 years ago. According to your logic, any CPU from 10 years ago with 3Ghz will perform better than a 2.8 from the current age. This simply isn't true. Speed is not the holy grail. New CPU's have different technologies that will perform tasks faster on the same speed because they have different functions and different instruction sets. You could compare it to a 2.0 liter engine in car A, and a 1.8 liter 16 valve big fuckoff turbo engine in car b. The 1.8 will run circles around the 2.0 liter because it's equipped with extra features.

[gak]

It's completely true, but hey, lately is the season for making dreadful mistakes...........

See "die hard"