Ryzen "2" ?

[gavinh87]

So the 8700k is 8% faster than the 1600 in single core at the same speed and 1.2% higher in multi core at the same speed.

That's a lot closer than some would have you believe and quite remarkable for a chip that cost's half as much.

Well, factor in the actual stock speeds and things look a little different but yes, decent performance when using ALL threads including SMT. Shame that many games don't use more than 8 threads.

 

[humbug]

Thats not 720P All you are testing there is the GPU.

So this is what you wanted?

 

[4K8KW10]

Interesting that you have higher mins.

Old school single threaded engine results

So this is what you wanted?

 

[Martini1991]

So the 8700k is 8% faster than the 1600 in single core at the same speed and 1.2% higher in multi core at the same speed.

That's a lot closer than some would have you believe and quite remarkable for a chip that cost's half as much.

It's the clocks that ultimately hold ryzen back. Unless Intel release something special then AMD could ultimately be the CPU king in a few years

 

[humbug]

Interesting that you have higher mins.

Oh yeah ^^^ not that this matters. its all academic.

 

[Martini1991]

Interesting that you have higher mins.

Not really. Heaven is a bit iffy when it comes to minimums. If the minimums were sustained then fair enough

 

[Doobedoo]

Well, factor in the actual stock speeds and things look a little different but yes, decent performance when using ALL threads including SMT. Shame that many games don't use more than 8 threads.

So it's a bit of a myth that Intel holds a meaningful IPC advantage just a clock speed advantage really.

 

[Martini1991]

So it's a bit of a myth that Intel holds a meaningful IPC advantage just a clock speed advantage really.

8% is pretty meaningful.
Combined with higher clocks you can easy end up with an extra 20% performance per core

 

[gavinh87]

@humbug yes

So it's a bit of a myth that Intel holds a meaningful IPC advantage just a clock speed advantage really.

Well in cinebench yes. It's important to note that ipc cannot be measured over 1 test. As we've just shown, the scores are quite close in cinebench but they don't transfer to heaven.

 

[gavinh87]

8% is pretty meaningful.
Combined with higher clocks you can easy end up with an extra 20% performance per core

Indeed, this is a gpu upgrade for some!
It's not without cost though.

 

[Doobedoo]

8% is pretty meaningful.
Combined with higher clocks you can easy end up with an extra 20% performance per core

Not really 8% is still quite small on a clock for clock basis. We're talking about an 8th generation product vs a 1st generation product after all.

Would we expect that difference to be even smaller with the new Ryzen line up?

 

[humbug]

@humbug yes



Well in cinebench yes. It's important to note that ipc cannot be measured over 1 test. As we've just shown, the scores are quite close in cinebench but they don't transfer to heaven.

No, You're quite right they do not ^^^ to illustrate that its obviously why you asked me to make a run of it with everything turned down and off at 720P

8% is pretty meaningful.
Combined with higher clocks you can easy end up with an extra 20% performance per core

Yup, i agree 8% does matter, especially when Intel also has the clocks.

But should those clocks close up more soon then.... umm.

 

[Doobedoo]

@humbug yes



Well in cinebench yes. It's important to note that ipc cannot be measured over 1 test. As we've just shown, the scores are quite close in cinebench but they don't transfer to heaven.

That is very true so you boy's should get to work and do some more testing. Chop, chop

 

[gavinh87]

No, You're quite right they do not ^^^ to illustrate that its obviously why you asked me to make a run of it with everything turned down and off at 720P



Yup, i agree 8% does matter, especially when Intel also has the clocks.

But should those clocks close up more soon then.... umm.

It was more to show how these CPU's will fare with future GPUs. Back in the good ol' days this was how it was done.

That is very true so you boy's should get to work and do some more testing. Chop, chop

Willing to run anything that doesn't cost money Unless ofc we already have the same game.

 

[Doobedoo]

It was more to show how these CPU's will fare with future GPUs. Back in the good ol' days this was how it was done.

I prefer to see testing done in real world scenarios to be honest. People are using them in 1080p, 1440p or 4k so that is how they should be tested. If it put the load on the Gpu then great as that is likely what the end user will also see when they pair it with that particular gpu.

 

[gavinh87]

I prefer to see testing done in real world scenarios to be honest. People are using them in 1080p, 1440p or 4k so that is how they should be tested. If it put the load on the Gpu then great as that is likely what the end user will also see when they pair it with that particular gpu.

They are plastered over the net already, no fun in that!
In all seriousness though, low res testing used to be done to see how well they would age.

 

[Martini1991]

Not really 8% is still quite small on a clock for clock basis. We're talking about an 8th generation product vs a 1st generation product after all.

Would we expect that difference to be even smaller with the new Ryzen line up?

The whole generation naming scheme shouldn't be used that way really.

Next ryzen line up should clock higher, i look forward to switching out my 1700 for an upgrade

 

[Doobedoo]

They are plastered over the net already, no fun in that!
In all seriousness though, low res testing used to be done to see how well they would age.

Not so much on a clock for clock basis as we have been discussing.
Used possibly being the key word. 1080p minimum in my opinion forget about making assumptions based on products that we know little about let's focus on what we do have first.

 

[Doobedoo]

The whole generation naming scheme shouldn't be used that way really.

Next ryzen line up should clock higher, i look forward to switching out my 1700 for an upgrade

Agreed 7nm Zen is the next gen imo but the 12nm ones will be better by some margin at least. but the point being that Intel have had a long time to squeeze that amount of performance from coffee lake, that is why 8% single and 1.2% multi seems quite small, common perception would probably be a bit more perhaps.

 

[gavinh87]

Not so much on a clock for clock basis as we have been discussing.
Used possibly being the key word. 1080p minimum in my opinion forget about making assumptions based on products that we know little about let's focus on what we do have first.

Tbh even at 1080p minimum on some games is enough to make our 1070 the limiting factor. I'd lean towards your opinion if we both had ti's!
What I have shown in another thread is that the 1600 can and is limiting a gtx 1080 at 1080p. Thus my comments about the fear I have for the lifespan of 1st gen ryzen.