Top end hardware - poor results

[lownslo 592]

8 minutes ago, vgbaron said:

Telling someone their opinion does not matter is a tad snobbish IMHO.

Curious where that happened in the thread?

Agreed that open discussions are important, but claims should be substantiated... at the very least folks should post supporting info along with their claims.  Then more of us can learn more.

Greg

[AnkH 2,264]

You see, as a scientifically educated person I still do not agree. Yes, if you talk about shoes, colours, taste and all this stuff, you can have an "opinion" and a discussion can be very interesting and fun. But the post was a clear cut claim that turning off HT did result in an increase of the FPS from 25 to more than 60, and no, I stick to this, this is nothing that you can call "opinion". It is a statement about a fact and if so, the one giving the statement should be able to proof it. If this is snobbish, well then, it is. Today's world is full of so called "opinions" and people tend more and more to use this "opinion" word for their sometimes mislead minds. Example:

[AnkH 2,264]

Ok, agree, slightly OT and exaggerated ;-)

[vgbaron 3,620]

51 minutes ago, lownslo said:

Agreed that open discussions are important, but claims should be substantiated... at the very least folks should post supporting info along with their claims.  Then more of us can learn more.

Greg

"should" be not "must"be. We are not publishing a paper here. Yes, it would be beneficial if a poster had charts and all sorts of bells nd whistles but then, those that disagree with the findings would get into an endless discussion on inaccurate testing methods.

Yes, it would be ideal if one posted a "finding" with substantiated backup and all sorts of empirical data but in all my years on these forums I can count on one hand, the number of people with the ability to do that clearly and correctly.

Meanwhile we deal in "this works for me". These forums are a melting pot of information, a gathering of knowledge from all sorts and posts should be accepted for exactly what they are - an individual opinion supported minimally only by the posters statement.

Now how about we get back to the topic at hand....

 

Vic

[Lenny777 299]

Where's the OP for this thread?A number of responses but no reciprocation. 

[TechguyMaxC 516]

3 hours ago, Lenny777 said:

Where's the OP for this thread?A number of responses but no reciprocation. 

Hasn't been online since Tuesday evening according to his profile.

[giantkillerrobot 1]

On 7/19/2018 at 3:07 PM, lownslo said:

Proof please.  Scientific proof.

OK, Proof:

https://imgur.com/a/BztILT9 

P3D, MILVIZ DHC-2, ORBX everything I could find for the PNW,   Active Sky and ASCA.

Saved scenario, HT on and off after a reboot. HT on yields 25-55FPS in flight, while HT OFF yields a consistent 55+ FPS, all other considerations being equal.  For my system HT has a quantifiable impact on performance, contrary to AnkH's opinion, it is not "nonsense".

 

This is a reasonably easy, non-destructive configuration change the OP could try that might have significant bang-for-the-buck. I"m a little confused why it has touched off a firestorm from AnkH and lownslo.

 

If you'd like to dig into how HT works Intel's tech docs are at:

https://software.intel.com/en-us/articles/introduction-to-hyper-threading-technology

 

Wikipedia has a pretty good explanation of the tech at:

https://en.wikipedia.org/wiki/Hyper-threading

 

[TechguyMaxC 516]

12 hours ago, giantkillerrobot said:

OK, Proof:

https://imgur.com/a/BztILT9 

P3D, MILVIZ DHC-2, ORBX everything I could find for the PNW,   Active Sky and ASCA.

Saved scenario, HT on and off after a reboot. HT on yields 25-55FPS in flight, while HT OFF yields a consistent 55+ FPS, all other considerations being equal.  For my system HT has a quantifiable impact on performance, contrary to AnkH's opinion, it is not "nonsense".

 

This is a reasonably easy, non-destructive configuration change the OP could try that might have significant bang-for-the-buck. I"m a little confused why it has touched off a firestorm from AnkH and lownslo.

 

If you'd like to dig into how HT works Intel's tech docs are at:

https://software.intel.com/en-us/articles/introduction-to-hyper-threading-technology

 

Wikipedia has a pretty good explanation of the tech at:

https://en.wikipedia.org/wiki/Hyper-threading

 

That's not how this works.

2 screenshots of a single point in time don't demonstrate anything, certainly not with the in-game framerate counter with its horribly low polling rate.

If you want to convince anyone you'll need to:

1) post a video

2) use a 3rd party *real-time* FPS measurement tool (such as FRAPS)

3) log the data

Something like this: 

Of course, you'll have to develop the testing scenario yourself but it sounds like you already have one in mind so really, this should be quite trivial for you.  

[Bob Scott 7,720]

14 hours ago, giantkillerrobot said:

OK, Proof:

https://imgur.com/a/BztILT9 

P3D, MILVIZ DHC-2, ORBX everything I could find for the PNW,   Active Sky and ASCA.

Saved scenario, HT on and off after a reboot. HT on yields 25-55FPS in flight, while HT OFF yields a consistent 55+ FPS, all other considerations being equal.  For my system HT has a quantifiable impact on performance, contrary to AnkH's opinion, it is not "nonsense". 

Well, not "proof," but demonstration of a measurable impact under one set of conditions.

I've found that HT On, with an AM=245 (hex F5, binary 11110101) that restricts P3D to one logical processor on Core 0, and (of lesser impact) on Core 1 not only does not exhibit any kind of slowdown, but it runs more smoothly than no AM with all cores enabled on a quad core 7700K.  I constrain frame rates to 30 by using VSync and a 30Hz 4K TV, so frame rate impacts are N/A (both configs hold 30fps)...I look for smoothness.  It worked that way on my 4790K (now relegated to driving sims and FPS games) and now on my 7700K dedicated P3D machine.  I don't offer that as "proof" of anything, but a countervailing observation when folks claim that no AM and HT Off is a look-no-further one-size-fits-all best solution.

Regards

 

[Bert Pieke 7,585]

As was discussed above,  "HT off" works nicely for a six core CPU (8700).

For a Quad, I would agree, a combination of "HT on" and affinity mask set is the better way..

[giantkillerrobot 1]

Well, I guess I’ll leave it as demonstration of measurable impact under one set of conditions, consistent with theory and documentation. 

 

Have a a good weekend flying. 

[ieydis 51]

47 minutes ago, Bert Pieke said:

As was discussed above,  "HT off" works nicely for a six core CPU (8700).

For a Quad, I would agree, a combination of "HT on" and affinity mask set is the better way..

Completely agree with Bert.  Wasn't a believer until I tried it.  8700K definitely better with HT off.  This is for P3D.  Any other more "modern" application or game, all bets are off.  P3D though is where I spend majority time on my computer so it's all that matters to me.

😁

[GSalden 4,657]

My 7940X :

No HT , no AM 

core 0 @ 4.8 Ghz

core 1 @ 4.7 Ghz

core 2 - 13 @ 4.4 Ghz

That is the best for my system..

- - - 

Regarding the poster that mentioned that HT off ( 55 fps ) compared  to HT on ( 25  fps ) gave way better fps  : probably the extra heat throttles down the cpu...

- - - 

Everyone has to find his/hers best system settings and best Fs settings to get the best out of it..

 

[W2DR 1,860]

This isn't rocket science work. The problem is the CPU. 3.3 GHz isn't even close to "top end". The only thing that will increase the frame rate is a faster CPU (as in 4+ GHz) .....just my humble opinion..........Doug

[Litmoose 90]

This CPU has a baseclock of3.3Ghz. It boosts to 4.3Ghz(and 4.5ghz on 2 cores) so its not that bad. I'm guessing what ever the issue is, it probably isn't down to the clock speed of his CPU.

Edited August 1, 2018 by Litmoose