Is Coffee lake worth the upgrade for flight sim?

[Sethos 4,294]

1 minute ago, Dirk98 said:

Question all you want, Sethos

Thank you.

[wizzards 99]

3 hours ago, DJJose said:

This is ludicrous. By now I should be able to OC a new CPU to 5gz with the press of a button and not have to worry about temps.

I could not agree more unfortunately Intel does not seem to share the same view of things,i wonder how we can make them see sense

Peter

[SpiritFlyer 987]

9 hours ago, Sethos1988 said:

The 8700K is going to be the best CPU going forward. It just hits a perfect trifecta between price, single-core speed and the amount of cores. Though as a few people have mentioned, I would highly recommend getting it delidded, as that's its biggest weakness, like with all new Intel CPUs. Personally buying mine pre-delidded from Caseking, that also means you retain the 2 year warranty through them, instead of pissing it out the window and having to do the work yourself. 

As for your question about the 8700K being better than the 7700K? In terms of raw IPC numbers, probably not, as they about equally fast. In that scenario, the benefit come from either being able to offload running apps onto the other cores, generally more overhead for other tasks and just being that tad more future proof. Considering the 8700K will become the main consumer chip going forward, buying a 7700K now seems daft unless it gets thrown in your direction for practically no money.

 

8 hours ago, westman said:

Agree,  spot  on 

I agree completely.

Even though it is perhaps too early to render a final judgment as hard evidence is still a bit thin, to purchase a 7700K right now would seem short-sighted indeed.

Kind regards,

[787flyer 633]

 

6 minutes ago, SpiritFlyer said:

Even though it is perhaps too early to render a final judgment as hard evidence is still a bit thin, to purchase a 7700K right now would seem short-sighted indeed.

 

Re buying the 7700K, do you think the price will tumble based on the 8700K coming out? How much do you think the price would depreciate for those of us on a limited budget to upgrade? I am on a 2600K right now OCed at 4.2 Ghz. Then of course there's the price of the motherboard to consider. Not sure what the price differential is between the Z270 and Z370 right now. If the differential overall is relatively small(i.e.< $100) I may go for an 8700K.

Regards
Mark

[SpiritFlyer 987]

4 hours ago, Dirk98 said:

Plus it's much larger CPU cache size (than 8700k's) may have played the main role in 85% FPS increase that Rob reported here recently in the neighbouring thread. And most probably due to the cache size in the first place.

 

4 hours ago, Sethos1988 said:

or maybe people just need to take that number with a truckload of salt. A larger cache will not give you an 85% boost.

 

3 hours ago, carlito777 said:

+1. That claim of a 85% increase in fps was ridiculous. There is absolutely no way to explain that, therefore I would highly recommend considering it to be another one of those "hey, my fps have trippled" stories...

 

3 hours ago, Dirk98 said:

Unlike many, Rob always gives proof for his findings and conclusions. And also unlike many, he has a reputation for that. Based on my experience I have no reason to doubt them, unlike your assumptions, Sethos ))

 

Dirk

I've seen where Rob was mistaken before, and maybe even dismissive when corrected, but I have never seen him dishonest nor anything but meticulous in his testing documentation.

If Rob said he got an 85% FPS increase with that particular CPU, he did. That does not mean that the production versions will equal that kind of performance, but if present testing produces similar results, that could amend all our present conclusions.

Occupational hazard: We may be obsessive in our pursuit of computational excellence, but let's remember that we are people that work better under a load of kindness, not units. 

kind regards,

 

[Sethos 4,294]

18 minutes ago, SpiritFlyer said:

I've seen where Rob was mistaken before, and maybe even dismissive when corrected, but I have never seen him dishonest nor anything but meticulous in his testing documentation.

If Rob said he got an 85% FPS increase with that particular CPU, he did. That does not mean that the production versions will equal that kind of performance, but if present testing produces similar results, that could amend all our present conclusions.

Occupational hazard: We may be obsessive in our pursuit of computational excellence, but let's remember that we are people that work better under a load of kindness, not units. 

kind regards,

 

I don't think anyone is calling his results a lie, as anything he manufactured them. Just calling them out to be questionable on the grounds there are variables unrelated to the sheer upgrade in performance that causes this. I can already tell from the fact the he has a 5960x, which I also do, that I rarely to never see my FPS go below 28. So his starting point already seems questionably low making any improvements or even a normalisation, seem like a huge improvement. There's so many factors in testing and benchmarking that could skew the results, sometimes badly, that you really need to be critical and question 'magical' results, as an 85% increase is when the IPC gains aren't even close to that. 

What you also need to watch out for is taking one, single source as gospel. Usually when you benchmark, especially when the results are too good to be true, you re-test, try other machines, try other hardware and ultimately, have other completely different sources to root out any anomalies. Taking ONE source, with highly questionable results, who isn't an actual benchmarking outlet to begin and then just running with that is silly. That's where you apply common sense and critical thinking, question the result and wait for more sources to actually verify it.

If you ever saw a single guy do some benchmarking on Intel's latest chip and it was 80% better than EVERYTHING else on the market, of course you'd question the result until it could be verified. Why on God's green earth would you just say "Yes I trust that guy, everyone questioning it is a hater and has no credibility". People here put an awful lot of faith in a single guy, a single set of results, anecdotes and good stories. Same reason why all the "I rubbed the side of my case and gained 300FPS!" stories gain so much traction every time, because people desperately want more performance out of their machines. 

[Dirk98 354]

6 minutes ago, Sethos1988 said:

I don't think anyone is calling his results a lie......

Whatever.

 

Rob' results are right until proven wrong (chew on that).

And Rob's reputation matters.

Dirk.

[carlito777 712]

1 hour ago, Dirk98 said:

Whatever.

 

Rob' results are right until proven wrong (chew on that).

And Rob's reputation matters.

Dirk.

He is really your superhero, is he?

I see no point in discussing this any further. Sethos has given a very clear explanation why he and I do not believe in Rob's results. But feel free to believe in anything you want. We live in a free world after all. For me, his findings are not relevant anyways as I would never spent this ridiculous amount of money on that kind of equipment just for a simming PC. I believe the 8700K will be the way to go as it offers great bang for the buck (if not priced significantly higher than the 7700K). But to each his own opinion.

[SpiritFlyer 987]

5 minutes ago, carlito777 said:

I believe the 8700K will be the way to go as it offers great bang of the buck (if not priced significantly higher than the 7700K).

I'm sure you are right and all the evidence, such as it is, points to it being an exceptional value. Looking forward to what is said when the NDAs run out very soon. There is a lot of excitement behind the scenes, I know that much. We will know a lot more sooner than later. These are great times for the advancement of FS, on all fronts!

Kind regards,

[Dirk98 354]

43 minutes ago, carlito777 said:

I believe the 8700K will be the way to go as it offers great bang for the buck (if not priced significantly higher than the 7700K). But to each his own opinion.

And I believe that the serious cache size difference (19Mb vs 11Mb on 8700K) may become the winning factor at the same 4.8-5.0Ghz freq easily achievable on both 7900X and 8700K CPUs (delidded). I'm sure end of November we'll have all the comparison data. And yes, 19Mb cache bites price-wise, no magic shortcuts in our hobby.  

Dirk.

[westman 257]

cashe / cores is  relevant,  check 7980X all I9 have the same amount of cash per core.

[carlito777 712]

40 minutes ago, Dirk98 said:

I'm sure end of November we'll have all the comparison data.

Fair enough. Let‘s just wait and see. Time will tell. 

[Sethos 4,294]

Well, the 8700k is releasing tomorrow officially. Amazon already had a few in stock today. So hopefully we'll get out data a lot sooner. 

[carlito777 712]

Here you go:

http://www.guru3d.com/articles-pages/intel-core-i7-8700k-processor-review,1.html

Looking good. Especially the potential to overclock to 5GHz or even higher. 

[vortex681 806]

20 minutes ago, carlito777 said:

Looking good. Especially the potential to overclock to 5GHz or even higher. 

But for single threaded performance, which is what counts at present for the current batch of flight sims, there's not much difference when compared to the earlier generation CPUs. On page 7 in the link, there's minimal difference between the single-thread IPC of the 6700k and the 8700k (and it's identical to the 7600k). What's more disappointing is that the single core CPU-Z benchmark score for the 8700k is no better than the old 4790k! Admittedly the multi-core performance of the 8700k is much more impressive but that's not going to make much difference for flight sims.