Intel i9 9900k to hit 5GHZ - Do you have one Rob?!

[Ted Striker 338]

1 hour ago, SteveW said:

...Noel, this is what you do:

Steve,

I appreciate you explaining that test process. However, for someone with a four core trying to decide whether to get a 6, 8, or 10 core it is not very helpful as they do not yet have the hardware to test. What have you found to be the optimum number cores running with HT and without HT?

Thanks,

Ted

[Noel 4,501]

1 hour ago, SteveW said:

...Noel, this is what you do:

 

You know the first process described by P3D is monolithic, running the user input and rendering. (So with HT enabled) it's obvious you need an AM with one of the LPs unallocated on the first core you allocate the sim. Get a scenario loading a few times and measure how long it takes with a stopwatch. Now add pairs of LPs on subsequent cores and each time measure the time loading the same scenario - let it load once at least before a timed run. Eventually the scenario will not load any more quickly no matter how many cores allocated - this is too many. On a pushed overclock drop one or two LPs. on a cool system leave the LPs at that count.

Yes, you set me up w/ an AM to get one LP unallocated for the monolithic process.   Now I'm not the first person in the many year history of multi cores and P3D and surely you've done this experiment previously.  Do you not have an educated guess for where you would think the sweet spot is for a SB-E chip?  

I'll try changing the AM to revert LPs into non-split cores, one core at a time, and see if I can ascertain this issue.

[Noel 4,501]

13 minutes ago, Ted Striker said:

Steve,

I appreciate you explaining that test process. However, for someone with a four core trying to decide whether to get a 6, 8, or 10 core it is not very helpful as they do not yet have the hardware to test. What have you found to be the optimum number cores running with HT and without HT?

Thanks,

Ted

Yes Ted, yes!  Surely SteveW has an informed opinion about this, and even though it will vary according to each users' demand one should get a sense w/ about how many cores/LPs are optimal.

[SteveW 1,792]

41 minutes ago, Ted Striker said:

Steve,

I appreciate you explaining that test process. However, for someone with a four core trying to decide whether to get a 6, 8, or 10 core it is not very helpful as they do not yet have the hardware to test. What have you found to be the optimum number cores running with HT and without HT?

Thanks,

Ted

A waste of everyone's time guessing - each system is unique, the hardware differences, the monitor differences, the demands and requirements of the user. 

[shivers9 1,174]

2 hours ago, Ted Striker said:

Steve,

I appreciate you explaining that test process. However, for someone with a four core trying to decide whether to get a 6, 8, or 10 core it is not very helpful as they do not yet have the hardware to test. What have you found to be the optimum number cores running with HT and without HT?

Thanks,

Ted

These discussions are interesting and I do very much always appreciate guys like Steve who try and teach us dummies about computers.😀 After many years of flight sims I just find that there are so many things that vary, it just makes me kind of skip over that subject when I see AM. For us mortals the idea of sucking that last once of performance out of our systems for the most part don't happen. We just have to live with the best advice I have seen. Buy the biggest, fastest CPU, GPU, Memory and cooling that we can fit into our budget (Lie to your wife if you have to) and then if the house lights dim when you fly into NYC and your system does not crash then you are tuned and tweaked!

Edited August 3, 2018 by shivers9
Duh

[jabloomf1230 2,445]

3 hours ago, martin-w said:

You will find zero evidence of the Noctua damaging motherboards

https://pcpartpicker.com/forums/topic/270785-noctua-nh-d15-weight

That's why they make retention straps and heavy backplates. It could happen, but like Rob said, who cares about a 0.00002% chance?

[Noel 4,501]

2 hours ago, jabloomf1230 said:

https://pcpartpicker.com/forums/topic/270785-noctua-nh-d15-weight

That's why they make retention straps and heavy backplates. It could happen, but like Rob said, who cares about a 0.00002% chance?

I think it's going to happen 100% of the time if the case is jarred hard enough!  Just intuitively I put a strap on mine just because it really LOOKS like the weight of it could tweak things in a bad way.  But alas after almost 6y and moving my PC case now and again, just inside the room, no probleme.   The other possibility is just orient the case flat it's possible to do.

[Ted Striker 338]

4 hours ago, SteveW said:

A waste of everyone's time guessing

I understand everyone's system is unique. I was just curious what you found with your system. I assume by your response about guessing that you haven't run the test on your system.

 

3 hours ago, shivers9 said:

Buy the biggest, fastest CPU, GPU, Memory and cooling that we can fit into our budget

That used to be easy to determine. In these days of multi core CPUs it is not clear that the biggest baddest most expensive 9980X with the most cores is the best for P3D as an 8700K with 6 cores that can clock faster may actually the better way to go. Both Rob and Westman have indicated that the faster clocking CPU may be the best for P3D. I was just curious as to whether Steve had run the tests on his system and what his conclusions were.

Ted

[SteveW 1,792]

26 minutes ago, Ted Striker said:

I understand everyone's system is unique. I was just curious what you found with your system. I assume by your response about guessing that you haven't run the test on your system.

 

That's  stupid Ted.

[SteveW 1,792]

4 hours ago, shivers9 said:

These discussions are interesting and I do very much always appreciate guys like Steve who try and teach us dummies about computers.😀 After many years of flight sims I just find that there are so many things that vary, it just makes me kind of skip over that subject when I see AM. For us mortals the idea of sucking that last once of performance out of our systems for the most part don't happen. We just have to live with the best advice I have seen. Buy the biggest, fastest CPU, GPU, Memory and cooling that we can fit into our budget (Lie to your wife if you have to) and then if the house lights dim when you fly into NYC and your system does not crash then you are tuned and tweaked!

The fact is that my post was responding to a Q. how many cores do I need? and to do that you do testing which requires an AM or how else could the test be made? So it is that P3D works fine without an AM - we are not discussing that. Seems like it fell on deaf ears.

Edited August 3, 2018 by SteveW

[shivers9 1,174]

Wow....when was the last time you guys had a laugh or saw the sun shine. Really....calling people stupid?

[Noel 4,501]

1 hour ago, SteveW said:

The fact is that my post was responding to a Q. how many cores do I need? and to do that you do testing...

Well for me, for my next purchase I'm trying to decide between a CPU that offers HT like I9-9900K and one that doesn't like I7-9700K.  Pretty hard to test since I don't have a test bed with those CPUs, just my old SB-E chip.  Or, is it possible 10 or 12 cores would be better than 8, HT or not?  Since I'm not particularly impressed by going from 4.7 to 5.0ghz, a meager +6%, the question is valid for me.  IOW, I still don't buy the argument that less cores for a few more tenths of a gHz makes sense UNLESS there is truly no value in having all channels open for those parallel processes, and indeed this is what we're trying to get at..pre-purchase! 

Edited August 3, 2018 by Noel

[Ted Striker 338]

1 hour ago, SteveW said:

That's  stupid Ted.

wow....all I was asking out of curiosity is what you found optimal on your system. That response ends my interest in this discussion.

Ted

[SteveW 1,792]

1 hour ago, shivers9 said:

Wow....when was the last time you guys had a laugh or saw the sun shine. Really....calling people stupid?

I'm not calling you stupid - I'm saying it would have been stupid to suggest a technique without having tested it myself....

[SteveW 1,792]

...I read back through all that Ted and it seems pretty clear my agenda was simply to answer your Q. about how to determine the number of cores - I answered courteously and with material worth observing. There's no optimal system here I'm developing so why not accept that I can't answer a plethora of questions like that - I give you  and the other guys the means to find out yourselves.