Install Steam
login
|
language
简体中文 (Simplified Chinese)
繁體中文 (Traditional Chinese)
日本語 (Japanese)
한국어 (Korean)
ไทย (Thai)
Български (Bulgarian)
Čeština (Czech)
Dansk (Danish)
Deutsch (German)
Español - España (Spanish - Spain)
Español - Latinoamérica (Spanish - Latin America)
Ελληνικά (Greek)
Français (French)
Italiano (Italian)
Bahasa Indonesia (Indonesian)
Magyar (Hungarian)
Nederlands (Dutch)
Norsk (Norwegian)
Polski (Polish)
Português (Portuguese - Portugal)
Português - Brasil (Portuguese - Brazil)
Română (Romanian)
Русский (Russian)
Suomi (Finnish)
Svenska (Swedish)
Türkçe (Turkish)
Tiếng Việt (Vietnamese)
Українська (Ukrainian)
Report a translation problem
Discussions Rules and Guidelines
Report this post
@quintencenshel and BRRM
it seems your apu is a notebook modell. AMD CPU's are always slower than intel cpus. At the moment the intel haswells have the highest ipc. Which means with an high clock the intel haswell cpus are the fastest.
Also you have an notebook cpu which are much slower than a desktop cpu and quad core amd is not the same as quad core intel.
A amd 7700 is also not really high end at best it's last generation mid range. Also a 7700 doesent exist it the series not the model number. It's a 7770 or 7790.
The clock of a CPU is not always so important. the ipc is also important (a lower clocked cpu can be faster than a higher clocked cpu). A fast haswell will always be faster than a amd cpu.
Also, the AMD APUs are cpus with integrated graphics. which are slower than a regular cpu + gpu.
amd cpu + amd/nvidia gpu or intel cpu + amd/nvidia gpu
Looking at your model it has a 4096 KB cache
My i5 has 6Mb
Now that may not sound like a big difference but performance wise it can be massive.
It's like having a fast car but severly restricting fuel flow to the engine, you aren't going to get the performance.
That GPU is better then nothing.
It can run desktop applications (even most videos) without issues. But only very low requirement games (console ports, games with simple visual style).
It is the replacement for the on-motherboard GPU's that used to be out there.
Now if you got one of the above CPU's and a dedicated GPU, things get tricky.
Unlike on-motherboard GPU's the CPU-GPU is NOT turned off. Both stay active. The Control Programm of the dedicated GPU decides wich programm uses wich Card.
Especially the NVIDIA programm is notorious to assign the wrong card:
http://steamcommunity.com/app/333950/discussions/6/617328415061110133/
I just had to override wich card was used for Star Citicen for Codd's sake!
Essentially AMD is settling at quad cores with 6 graphics processors. A key problem for the OP, as pointed out above, is the cache size on the A10, slowing down redundant fetch processes.
His max turbo speed is another issue, as well as wattage may also be constricting him.
Op, if the processor listing I linked to isn't your processor, then likely you have a lesser processor than the above.
The third strike against the OP's A10 is the memory speed, though DDR4 is showing up on mother boards, and is very expensive, I'll figure you've still got DDR3. That is a retail average use moderate speed ram, that you've moved far in distance, in comparison to a dedicated graphics card, from the cores, solely because the GPU is located on the main processor. Basically longer travel distance to slower RAM is a problem that is only compounded by the smaller cache.
Most graphics cards are 128 bit or larger in "data path" that means they can, in parallel, grab more information at a time, 4 times more than a regular CPU bus. They also have dedicated Video RAM on the card with a dedicated controller (traffic cop) for the management of that RAM. This problem is affecting AMD and Intel alike since they decided, I guess a decade back, to stop at 64 bit for some unknown reason, biting them in the butt how that they're putting graphics cores as well.
Your CPU is 64 bit, and I believe the FM2/FM2+ is also 64 bit, which means not only are you slowed down by 1/2 the ram speed, but the cores can't transfer memory at any faster than the clock rate times 64 bit chunks (in essence you're bottlenecked to 1/4 the data transfer rate graphically, as compared to a dedicated graphics card with DDR5 Memory and a 256 bit or higher data path). This exacerbates the smaller cache issue as well.
CPU is fine, and, that it's on a laptop I wouldn't play much, if at all, with overclocking, but if it's just a laptop CPU in a desktop, and you are about to build a new one, then have some fun and experiment some with this one. Hopefully you might have some bus controls that ramp up in accordance with the changes to the processor.
For mention to say anything is possible, I run an AM3+ 8120 (though I got one with the 6 mb cache and 95 watts somehow), 3.2 ghz stock, set at 3.8 ghz right now and turbos beyond that on its own (processor used to turbo to 4.0 ghz stock), still air cooled by one of those Coolermaster Evo Hyper 212 coolers, not even dual push pull fan set up either haha
Good luck OP and hope you get a system going that allows you to fully enjoy Space Engineers.
Edit: Forgot to mention that, also, since the CPUS and GPU are sharing that cache space, your CPU is being negatively affected by the game being more GPU intensive. Basically the cache fills with a variety of redundant CPU calls, then the GPU runs and the processor detects a number of reduntant calls here, flushes the cache and loads it with those, and with the next GPU instruction this occurs again, rinse and repeat of wasted cycles, again this is just a general explanation regarding the cache, as, even if the controller isn't doing a 100% flush or able to divide the cache in an efficient manner, those are clock cycles lost for the controller to come to that point of equilibrium, and, as you move through a game etc., that equilibrium is shattered over and over.
Exactly! All of these comments about a 2yr old system being a low performer is duh! You are the ONLY one that has addressed my question. The PhysX should be loading my GPU more than 4%, however they are still in DirectX9 on SE. Since they are talking about going DirectX11 I'm hoping they use it right. I hav e a program that lets me see the engines being used and what hardware they load. Problem is VRage is CPU load only and does not utilize SSE4 or higher functions. My main concern is currently the tax on my system is 2000s era programming. Do they intend to step up to modern hardware and utilize DirectX11 to it's fullest or are they going to bit and piece it like they are now with DirectX9. OpenGL would be a dream but these types (sandbox) of games can't seem to use it right either. There are a bunch of open source physics games online that balance the load between the calculations (CPU) and the rendering physics (GPU).
I'm about to build a new PC with AMD's new APU core. I prefer the APU's for power stepping, My computer doesn't pull the max wattage unless everything is loaded CPU/GPU. This is incredibly more power efficient then Intel's as they tend to max wattage at 50% load on either CPU/GPU. Not saying Intel is bad, but that it doesn't display the behavior I prefer. Yes, Intel benchmarks quicker. But seriously, having anything over 60 fps on a 5ms monitor is just a waste as the display does not refresh at that rate. And having my kids play games on my 4K 120Hz living room TV is not going to happen. Also my computers are primarily for my work so I'm only interested on if I'd have to go high-eng on my CPU cooling simple for a game.
No offense, but did you even read my concerns? I know all about AMD since it's A factor was created by a worker at Intel and they said it was crap so he bought it from them and started AMD. You obviosly are a text book user with no BIOS programming or kernel modifying ability. You are quoting non-dev options that I proved obsolete in the 4th grade. My laptop may say HP, but it was custom built and then custom programmed. Yes the brochure says switchable graphics, however I can also run dual since the MB has been modified. I buy manufactured laptops so you can see it may be junk, but I've also never had one stolen because of it's appearance. My CPU is not soldered to the board so I can swap at anytime. Look of HP's website for a DV6 CTO and it says they are, hence the CUSTOM part in my order. I talk directly with the builders so my definition of custom and what they put an the web are totally different.
And also when I say I'm designing a new computer, that means I've been on the phone with ASUS, Seagate, AMD, and G.Shill directly. There are no off-the-shelf parts in ANY of my system. You commented on DDR4? Have it in another PC. My video cards will also be custom creations. Even my hybrid drives are custom. My PROBLEM is SE bit and piecing DirectX9 so only to use 4% of by GPUs (yes, plural in my laptops case). Normal utilization of DirectX9/10/11 is to balace the calculation load between CPU and GPU. Physics calculations are CPU and Graphics rendering is GPU. This game DOES NOT work that way. They do NOT use SSE4 or higher CPU functions. See, I am running the game and looking at my diag screen on a second monitor.
This was my question, "Will SE continue to bitand piece DirectX11 functions as they are now with DirectX9?" There was NO comment about my systems performance other than my CPU load is high and my GPU load non-existant. As with Minecraft, KSP, and SE they are high-ending the lower CPU functions and ignoring the higher CPU(SSE4) and the GPU entirely. Which any programmer will tell you is NOT how DirectX9/10/11 are supposed to be used.
The offer to help must pertain to the problem at hand, not with what you read in PCMag while on the toilet.
First to the cooling, as I said I am still air cooled. Newegg often has these on sale and with free shipping, http://www.newegg.com/Product/Product.aspx?Item=N82E16835103099&cm_re=212_evo-_-35-103-099-_-Product (note I got mine for 20 USD w/free shipping, and make sure you get a nice 120mm fan since I know these fans on this thing burn out in 6 months, that's the drawback. 70 cfm or higher will be great and you can have 2 fans (that push/pull thing) on this). I don't know where you are or if you meant for a desktop, but this thing really has made the difference for me, to gain longevity and save my pennies for something better that I can still do to it what I am doing now to my 8120 haha
I'll figure they'll piecemeal the DirectX11 if they can (don't know if Microsoft is doing any sort of push to get rid of anything less, they do that sometimes), and yeah an OpenGL implementation would be great, however you might look at Keens website and info on their engine, I haven't in a long time but perhaps that could be useful in making suggestions regarding the openGL.
As to the division of processing power, as long as your gpu is on a 64 bit path, and with such a small cache, your potentials are limited. You might search for other utilities to adjust it but I am going to contend that your best bet is the fastest DDR3 you can get and overclock that. Make sure you've set up windows to use a thumb drive or SSD for max boost and VM, and get a great cooler. You might look into the pci/pci-e latency setting (if that's an option in cmos) it might be set for a high delay, i'd make it as low as possible as that would get rid of unnecessary delays in the system assuming the cores a PCI device to draw the graphics to.
That's about all I can suggest and I do wish you luck :)
I'm sorry, you appear to have come to all kinds of conclusions as though someone is trying to make an affront to you. I am not, was not. I offered what I can. That you took offense, I don't know why
I'll leave you to your certainties, but do not apologize for making an effort to help. God love ya.