My R7 is coming today and my GTX 1080 should be coming on Thursday, so once everything's set up I'll do some in-depth testing. I'll probably try the game at 1080p and 4k in a variety of parks, while overclocking the chip however far it can go on the stock cooler. I'll also try disabling cores and seeing how it affects performance, to get an idea how the 4 core and 6 core chips will compare to the 8 core chips.
Anyone testing planet coaster with new ryzen cpu?
- Thread starter smelly
- Start date
Very interesting to hear that there is zero difference between the 7700k and the 6950X which basically tells me exactly what I mentioned about what others had speculated that at this moment in time, PC might actually be hitting a game-engine bottleneck of some sort because there should easily be a massive difference between those 2 CPU loads in terms of producing raw FPS performance difference. It almost sounds like your not being honest with your findings compared to what others posted as their results with 10k guests on a i7 5820k CPU (which during the launch of PC was one of the more popular cpus for this game) on the planet coaster reddit and what people have commented on in the parks in the steam workshop. I don't have either CPUs though so I just have to take your word for it, but that seems highly unlikely that a 4 core cpu would be producing the same FPS as a 10-core i7. Highly Unlikely. Maybe at lower guest count perhaps yeah, I could see that, but once you start approaching over 5k guests, the difference shows.
As for the BF1 comment, no it doesn't utilize all cores properly because it is a game that is built on an engine that still favors clock speed over core-count, which is why Broadwell and Ivy bridge i7 4-cores perform so well or better than those CPUs with more cores, but I wasn't talking BF1 specifically, almost all reviewers are still using BF3 and BF4 as benchmarks which both don't utilize more than 4 cores properly, and favor clock speed over core count, like pretty much every game over the last decade has.
Back to PC though, only real performance benchmarks can tell the Ryzen story with this game which I'm still waiting to see really, there have been 2 people so far on these forums that had one, which both conflicted each-others benchmarks so still waiting to see if more people can report what they get.
Last edited:
I have a 5820K.. and have posted my fps here on certain parks... if you want me to check a different park I can.. At this point I'm thinking maybe my RX 470 is holding me back and need a better video card. People with GTX 1060's are getting better results than I am.. and the RX 470 should be slightly better than 1060.
Ive got the Benchmark Video done now. It will be available under the following Link. Sorry, no time for a detailed describtion at the Moment, have to go to work. [cry]
https://www.youtube.com/watch?v=qaFf46KPSvQ
https://www.youtube.com/watch?v=qaFf46KPSvQ
This right here is very suspect to me. While the RAM modules may be 3200, that is the max OC value of them. They run at 2133 natively, and as of right now, there aren't really any boards getting the RAM up to that, and those that get close are with some pretty serious bios tweaking. I have mine running at 2933 14-14-14-30 timings. I am also running a 1700 with all cores at 3.8 ghz and have not fully tested all the parks yet...got my 1080 ti arriving Thursday, and am tearing it all down right now for my water loop.
Last edited:
The difference isn't massively improved frames, it is consistent smoother frame pacing. The more cores in PC seem to from my findings and others since Alpha when we tested a number of parks was always that there was little to no real difference in FPS. Now this may be because we are all bottlenecking our GPU prior though as most were on 980 & 980 Ti's because the 10 series hadn't landed and with the 1080Ti I have seen all 4 core CPU bottleneck at 1080 & 1440p in reviews so far.
Only 4K did it become GPU bound again. And I should say I am playing at 4K when I test because that is what I game at. If I played on low 720p then I would likely see a difference but we are still talking going from 60fps to 65fps. The bottleneck itself seems to be a number of cross talk points from the CPU to GPU from what I can tell. That comes down to how the engine is moving the data.
BF1 multiplayer utilises all the cores and threads even on a 6950X. It tops all the charts from all reviews doing such. Just have a google, they have been out there for a month now showing that the 7700K and lower 4-cores are all beat by the Ryzen & Enthusiast chips in that. Ignore the single player as that is coded for 4 core naturally. But the way multiplayer works . I have not seen a review in two years that use BF3 to compare so sorry but I think your info and what you are reporting is dated somewhat. I should also state they tested the game at 1080p with a titan X pascal and 16GB RAM.
If you pop across to overclockers forum there are a number of peopel running 16GB at 3200MHz just fine with a 3.9GHz overclock on a 1700.
There are a few that have actually got the latest Bios for the CH6 with RAM up to 3600MHz 16GB, the problem is they are not able to complete Prime so dropped it down but was doing fine in game. Don't be so quick to dismiss it because you are not achieving it. I would also like to point out that 3200MHz has been possible for over a week on latest Bios. Some have gone direct to manufacture to ask for Beta Bios which are not publicly released.
In regards to what you have, are you on T1 or T2 timings as it appears T1 makes more difference with looser timings than T2 with tighter timings at this point with current bios.
And I would also say that the max OC value is what it is at a set voltage level, they can often go higher if you get the volts up to 1.45V and to give an idea, max volts for DDR4 is around 1.5V so a lot of headroom. Also you can pick up RAM with speeds rated to 4200MHz at 1.4V
We don't know what the poster is using to achieve this and as every different board model even from the same company is very different I don't believe we can question what they are stating without evidence.
I would say also that I used to be a system builder for a well known online company and have built and overclocked over 1000 PC's with all sorts of hardware while a spotty kid but we are going back 7 years when I done it as a job. I do however still build a lot of custom machines and watercool them etc for private clients. So that is my experience to really test and bench. I got out of it because well money mostly, it doesn't pay great even at the top end and also I didn't find running benches after benches all that interesting after a while and just want to game for my enjoyment.
Last edited by a moderator:
Please use the multiquote option when replying to multiple posts.
I haz done witchcraft to mangle them all together.
I struggle to keep train of thought for more than one post at a time, especially when technical. In some cases I don't know I am going to reply to multiple people as I read something and hit reply with quote. Then I will continue to read and then reply to something next. I have not learnt to predict the future at this time so had no idea I was going to quote more than one person.
I also find it very difficult to read through multi quote replies so that is why I don't. I think I will just go back to lurking because this multi-quote is that frustrating to me. I have never used it on any forum before and will not start now as it is not needed in my opinion.
Edit: Just in regards to RAM this is what someone on the overclockers forum are managing at moment:
Last edited by a moderator:
I have spent a lot of time on various overclock forums for years
Yes, there are people hitting the 3200 on a few motherboards, but not even close to the majority, and what works for one person is not working for many others, so it is still quite the wild west when it comes to tweaking Ryzen, which is to be expected given it being an entirely new platform and all.I have 1T RAM and do plan on pushing my system a bit more because I know I have more headroom left to play with. So far, I have achieved my numbers without manually adjusting any of the voltages, and don't really plan on it until I am under water.
I think my point in questioning those numbers was from the information given, there was no indication of any tweaking or OC'ing given, or even setting XMP in bios, and typically when a person has to do that to achieve certain numbers, they say so. Granted, not everyone explains how they get the numbers they do, but there was no indication that anything was done, and one thing I am sure we can agree on is that nobody, no where, no way, no how has gotten 3200 straight out of the box.
Joël
Volunteer Moderator
Hey Curlyriff,
Thanks for your feedback. We don't mind if it happens on occasion, as long as you try to keep posting multiple posts after each other to a minimum. After all, we strive for a safe, friendly and enjoyable platform for everyone, including you.
I have spent a lot of time on various overclock forums for years
I have 1T RAM and do plan on pushing my system a bit more because I know I have more headroom left to play with. So far, I have achieved my numbers without manually adjusting any of the voltages, and don't really plan on it until I am under water.
I think my point in questioning those numbers was from the information given, there was no indication of any tweaking or OC'ing given, or even setting XMP in bios, and typically when a person has to do that to achieve certain numbers, they say so. Granted, not everyone explains how they get the numbers they do, but there was no indication that anything was done, and one thing I am sure we can agree on is that nobody, no where, no way, no how has gotten 3200 straight out of the box.
No problem. That is fair enough, however for instance I never go over how I get there unless asked because for the conversation/topic in question etc I don't feel it adds anything. Never mentioned majority or anything, just that it has been possible with the right board and RAM. Some just got lucky when getting the parts, others are working hard to reach there. I get all that but it to me still doesn't dismiss that it can be the case which is what you suggested in your initial reply.
Glad to hear your going well with 1T timing. Sounds like it's going reasonably well. I missed it but what board and Bios are you on?
I know at one point with the CH6 before they messed about with other things a few were using AMP/XMP profiles (same thing as I am sure you know but AMP is the specific for AMD and XMP is Intel specific technically). Incidentally where they are using bios actually from their Intel boards some are still named with XMP in the bios. (side note too all but just for people who may not be aware).
No worries
just hard at times. I have managed it here as only two posts to read so all good just gets frustrating on the longer threads, and replies that are paragraphs long. Not such an issue with a thank you though Normally, I don't go into detail about numbers either, but this is a more technical thread than most threads on this forum, and one that many non techies are going to read, and possibly make decisions and purchases based on what is stated here. While I don't think it is important to go into the nitty gritty specific details here, I do think it is important for people have realistic expectations, and that includes knowing that they will not have super duper top performance out of the box. Of course, they wouldn't with Intel at stock either, but we could go way into the weeds on the minutae that would not really mean much to anyone here
I have played around with the ASRock X370 Killer SLI/AC on 1.63 so far. I have an MSI Carbon ready to go as soon as I finish my scratch built case. I also have an MSI Tomahawk and an ASRock Taichi still in their boxes waiting. I am liking the ASRock bios so far. The stock one was was crap, but the updates have been very good, and it is clear that ASRock has matured in the UEFI department. I still have a 2500k running at 4.4 on an ASRock p67 board
Normally, I don't go into detail about numbers either, but this is a more technical thread than most threads on this forum, and one that many non techies are going to read, and possibly make decisions and purchases based on what is stated here. While I don't think it is important to go into the nitty gritty specific details here, I do think it is important for people have realistic expectations, and that includes knowing that they will not have super duper top performance out of the box. Of course, they wouldn't with Intel at stock either, but we could go way into the weeds on the minutae that would not really mean much to anyone here
I have played around with the ASRock X370 Killer SLI/AC on 1.63 so far. I have an MSI Carbon ready to go as soon as I finish my scratch built case. I also have an MSI Tomahawk and an ASRock Taichi still in their boxes waiting. I am liking the ASRock bios so far. The stock one was was crap, but the updates have been very good, and it is clear that ASRock has matured in the UEFI department. I still have a 2500k running at 4.4 on an ASRock p67 board
See I really like talking about all this but reading it through and thinking how many are going to be lost. Just for others reading then who maybe are not generally into the tech but this may help with the jargon being stated between us all.
With regards to RAM both AMD & Intel support RAM overclocking with pre-installed settings which for AMD are known as AMP and for Intel are known as XMP.
To use these settings you need to go into the Bios and select the profile that matches to your RAM chip. This often just needs to select enabled and it will select the one that matches your RAM.
Now with that in basics, currently AMD requires them to test all the boards and RAM chips and provide validation codes to the board manufacture to then implement them. This has meant that at the moment the Bios are very much still being updated with these and they are causing the issues the last few posts have been discussing.
Intel on the other hand don't require this so it is just down to the board manufacture to sort themselves and it works much better as it is a lot faster to implement.
This then becomes important as the R7 series from AMD are actually two 4 core cpu's glued together via a method AMD are calling infinity fabric. Now this is important as infinity fabric relies on the speed of RAM to make sure that connection is fast enough to work properly and not cause a bottleneck.
For every 20% increase in RAM speed we see a 10% increase in the Infinity fabric link speed.
The reason the issue is so obvious is that the bios for all different boards from even the same manufacture are at different levels and so we are not seeing consistent RAM speeds being attained on the same setups. What I have suggested though is that with relatively little tweaking now that at least one board, the Asus Crosshair 6. Now I don't particularly rate Asus and their products, I feel they are overpriced and normally gimmicky.
A round up of other motherboard manufacture
- ASRock are great though and certainly like to push things from previous experience.
- MSI I avoid as their CS is the worst, even more so than Asus now and that says a lot.
- Gigabyte seem to cut corners although overall I haven't had an issue they are heading the Asus way which isn't good.
- Trying to get a hold of Biostar is difficult but they genuinely seem to do their own thing with items such as the first to show their smaller ITX boards (literally motherboard size for people who don't follow tech).
With all of that said above and everything else so far I would say you will see in real world zero fps difference between most of these CPU's.
One thing I would add here to your great post is that given where mid range and higher CPU's, and GPU's are right now, aside from the odd duck gamers running 240hz monitors, probably the least talked about but by far the most common bottleneck on most systems today is...the monitor itself. With the most common being 1080p at 60 hz, that means the monitor can only display that many frames per second, regardless of what the CPU and GPU are pumping out.
Now with regards to how that translates with the R7 1700 and the 7700k, while the benchmarks have shown the 7700k will consistently get higher FPS, those same benchmarks have shown that the 7700k has a wider variation between it's mins and maxes than the Ryzen. What that means, and I have experienced it myself, and multiple reviews have said the same thing...Ryzen gives a smoother gaming experience.
I can confirm this. As i said in my Benchmark Video, the framerate was the same when i locked Planet Coaster to 8 Threads in the native Cores, but it felt more laggy. So Ive got the same framerate but a smoother Experience with 16 Threads Running.
Thanks and to go on with your view of monitors. You are indeed correct.
The Hz of a monitor tells the person how many frames will actually be displayed to the user so a 60Hz monitor shows 60fps literally. A 120Hz 120 fps is shown. These are maximum I should say as if you are only getting 40fps that is all the monitor will show.
And I will then go on to say that although the min, average and max are all lower than that of the 7700K for instance, they are in a smaller range which makes the game feel smoother because you are not visually seeing dips where frames are being dropped.
This gets a little more complex in that if a GPU is able to pump out a constant 120fps on a 60Hz monitor it will still appear smoother than something pumping out 120 with drops to 60fps. This is due to the way that frames are lost when sent to the monitor. Freesync and Gsync are now there to make this better. What they do is fill those drops that we see with the same frame repeated so that it looks smoother.
The reason we notice it is because we see it as stutter when the game drops frames. It is the reason older games that ran at 20-30fps feel smoother than some modern games that may run at 100fps but then have massive drops to the 30-40fps range at times.
Smoother to me is more important than higher overall fps. I have also seen first hand that Ryzen looks smoother due to it's smaller range of frame. I don't however know how it compares in honesty if you played with G-sync on both and then compared as I have not got a G-sync monitor to hand. Been waiting to see how Vega and Volta turn out before committing.
The reason and I stated before that on CPU's that have more cores etc than the usual i7 7700k for instance is because the game is loading all cores to 80% in the 7700K and it loads all the cores on say an i7 6900k to 60% because what it is able to do is distribute more processes across greater resources and thus you don't have the CPU spike as it does things which also causes issues with smoothness.
The following on with W10 gaming mode that is meant to be able to then take 2 cores and put them as allocated for purely windows and background tasks and then your other 4-8 cores (depending on chip R5 to 6950X) as your gaming cores which should make things even smoother because the CPU and Windows isn't trying to move things about to make space for one another constantly.
Edit: The reason I say real world and not CPU fps is that for 95% of people they are still GPU limited. Unless you have SLI 1080Ti right now and then an i7 6950X at 4.4GHz overclocked and then you have a different situation. That would then really show what things are like. That and then compare the 1800X at around 4.1/4.2GHz and the i7 7700K at 5.3GHz (they can be brought at that speed) and you will see what Ryzen compares like to the cream of the crop from Intel in terms of both 4 core and enthusiast 10 core.
You of course would need need real world builds with terms of M.2 SSD, secondary SSD, decent power supply and maximum support RAM so 64GB on the 6950X is likely the case if you are spending that on the CPU where as the 7700K and Ryzen are likely to be 16GB with the 7700K able to run 4000mMHz RAM at those speeds compared to 3200MHz.
I would also like to state that although as above MSI CS is awful it appears they have sorted out the memory on their latest bios so it is the select and enable the faster speeds for Ryzen. They have termed it A-XMP.
Last edited:
@Curlyriff thanks for the detailed posts on this thread, which really should be stickied to this forum -- and just about any other gaming forum out there as well. You've laid out a great deal of essential info concerning fps vs refresh rate and in the process have shown that the priorities of the fps cult are misguided. Thank you. Thank you. Thank you.
I would only mention that movies run at only 24 fps and as you've stated, it is above all stability, not a higher fps number, that creates smoothness and fluid movement.
I would only mention that movies run at only 24 fps and as you've stated, it is above all stability, not a higher fps number, that creates smoothness and fluid movement.
Alright, my Ryzen 1700 rig is put together and as a preliminary test, I loaded up a park with 4,600 guests (Elbinea on the workshop). I ran the game at the fastest speed possible while doing an aerial view of the park at 1080p low settings, to ensure there was no GPU bottleneck. When forcing the CPU to run with 4 cores and 8 threads, I DID see a performance drop: 13-14 FPS versus 15-16 FPS with all cores enabled. So while in this extremely CPU intensive scenario there is an improvement in performance, it's pretty minor.
I'll be doing more tests soon, hopefully incorporating more dynamic tests while tracking frametimes and minimum frame rates. I think someone here said they saw better minimum frame rates with more cores, so I'll be interested to see if the same is true for me.
I'll be doing more tests soon, hopefully incorporating more dynamic tests while tracking frametimes and minimum frame rates. I think someone here said they saw better minimum frame rates with more cores, so I'll be interested to see if the same is true for me.
Hi.
Please could any Ryzen owners test out my latest park and report back with their FPS (I'm getting around 22/23 paused and 15 running with i7-5930K and Titan XP - this is at 1440, with most settings high/ultra).
http://steamcommunity.com/sharedfiles/filedetails/?id=891694053&searchtext=century+park
Many thanks
Please could any Ryzen owners test out my latest park and report back with their FPS (I'm getting around 22/23 paused and 15 running with i7-5930K and Titan XP - this is at 1440, with most settings high/ultra).
http://steamcommunity.com/sharedfiles/filedetails/?id=891694053&searchtext=century+park
Many thanks
Last edited:
At Elbinea, i got around 26 to 29 FPS in CPU Limit on Elbinea with a Ryzen overclocked to 3.8GHz. Most of the 1700 CPUs are able to reach these clock speeds if you are interested in learning how to overclock your CPU. Ive just testet Planet Coaster limited to 8 Cores/8 Threds as a special test and i got more inconsistent Frametimes. I recommend to let it run on all 16 threds, it feels much more fluid.