Author Topic: Some thoughts on building a video gaming computer  (Read 1641 times)

0 Members and 1 Guest are viewing this topic.

Offline Electro FanTopic starter

  • Super Contributor
  • ***
  • Posts: 3320
Some thoughts on building a video gaming computer
« on: July 15, 2020, 02:17:08 am »
I recently helped build a gaming computer.  Here in no particular order are some thoughts/lessons learned:
1.   For games a key metric is Frames Per Second.  FPS can range from roughly 60-240 or more.
2.   The FPS you are going to achieve will depend on the refresh rate of the monitor, which in turn will be heavily dependent on the resolution.  The impact of FPS (and some related specs) is the subject of this youtube video:

3.   It turns out that about two-thirds of gaming is done at 1080P.
4.   Some games are written so as to relatively easily achieve high FPS and others are written in a way that makes high FPS more difficult to achieve.  Once you pick your reference game and game settings you can then use that game as a starting benchmark in designing your computer / selecting your components.
5.   In addition to having a monitor that can achieve 240 FPS (by supporting 240 Hz or better refresh rate), a/the key link in the chain is the GPU.  It is possible that a very high-end GPU (Nvidia 2080 Super, etc.) can output more than 240 FPS but the monitor will then be the liming factor.  Some monitors can support 280 Hz and therefore 280 FPS.  In some cases you can see even higher (than 280) FPS numbers displayed as a performance readout on the monitor but this represents what the GPU is outputting vs. what the monitor is doing.
6.   The CPU needs to be able to do it’s share of the work, but generally the bottleneck in terms of video gaming performance and price is the GPU.  (For a non-gaming PC used for business applications such as MS Office you can save a lot of $ by going with an entry level video card or using the motherboard's built-in graphics function.)
7.   Intel CPUs are holding their own in terms of performance when using relatively few cores and few threads but as application software is able to use relatively more cores and threads the AMD processors become increasingly competitive.  How much the number of cores and threads supported will impact game performance depends on how the game (or other application) was written.
8.   For some reason (probably economics and inertia), Intel has been somewhat slow to reduce die size; conversely AMD has been more rapid to market with smaller die sizes.
9.   One design aspect that AMD has a clear current (but probably temporary) lead is in the ability to support PCIe 4.0.  A PCIe 4.0 bus will deliver twice the bandwidth of a PCIe 3.0 bus.  One area where this is starting to become an advantage for AMD (and AMD motherboard makers, such as Asus and others) is in system storage.  NVME drives that support PCIe 4.0 are now available.  They can achieve sequential reads and writes of approximately 5 GBytes per second – which works out to be about 4 minutes to write a TeraByte.  This is somewhat gated by the controller chip functionality that appears to be ready for another step forward (maybe this summer) which will enable PCIe 4.0 NVME drives to achieve about 7 GBytes per second.  For games this is probably not a big deal as the game gets loaded from the drive into RAM and then the GPU becomes the primary workhorse – but if you want to make a screaming fast general purpose computer, NVME and PCIe 4.0 are the way to fly.  As a frame of reference, a Western Digital Black 7200 RPM SATA drive might achieve on the order of 270 MBytes per second (not even close to top of the line NVME).  Overall, depending on whether the data is being read/written sequentially or randomly and in what data sizes, NVME drives will perform 10-100 times faster than conventional electro-mechanical drives.  NVME might not provide the same very long-term data durability of conventional drives but for many/most users it will probably last as long as the initial computer it is supporting.
10.   Before selecting your RAM, check with the motherboard manufacturer’s Qualified Vendor List to see which specific makes and models of RAM have been tested for compatibility with the motherboard.
11.   When installing Windows 10 on a gaming PC hard drive (NVME, SSD, or conventional drive) be sure to select GPT vs MBR.  You want to confirm this before loading lots of application software.
12.   In order to squeeze the most out of a high-end gaming PC you need to have a big enough power supply to support the GPU and the rest of the components.  In addition to enough power, you need sufficient cooling.  An air-cooled PC can work fine but if you want to maximize your cooling (especially for the CPU) you want water cooling.  Water cooling comes in two flavors:  closed loop with pretty much off the shelf components, and custom do your own plumbing.  Closed loop off the shelf will get you the added performance with less work, cost, and risk versus custom plumbing.
13.   When building a gaming computer you can just go for performance or performance and aesthetics.  Aesthetics will probably take you into the realm of RGB lighting.  This can be fun or slightly more challenging and more expensive – depends on how you look at it.
14.   Selecting a case that will fit all the components and provide good airflow can be as important and as challenging as selecting the rest of the components.
15.   No doubt, much of this is subject to discussion and maybe some debate.  YMMV
16.   The intent of this post is create a thread where people can do some Q&A on system design, component trade-offs, etc.  Happy computer building!     
   
« Last Edit: July 15, 2020, 02:34:12 am by Electro Fan »
 

Online SiliconWizard

  • Super Contributor
  • ***
  • Posts: 15926
  • Country: fr
Re: Some thoughts on building a video gaming computer
« Reply #1 on: July 15, 2020, 02:43:44 pm »
8.   For some reason (probably economics and inertia), Intel has been somewhat slow to reduce die size; conversely AMD has been more rapid to market with smaller die sizes.

How so? Can you be more explicit on that?

Intel has usually been one step forward using the finer process nodes. If die size is not reduced significantly , it's because it's influenced obviously by the content, but also the number of pads (which is pretty large on a modern CPU.) What takes area as far as "content" goes are cores of course, but also cache memory which takes a lot of area. Reducing die size would basically mean reducing the number of cores per die, and/or cache memory. Or significantly simplifying the architecture, but that would probably be a major overhaul. What exactly do you see in that for the end-user?

I've seen recent AMD CPUs with many cores are actually made of several interconnected dies (basically a 32-core CPU is made of 4 8-core dies), is that what you mean? But what's the benefit for the end-user? And does Intel not do the same with their >= 16 core CPUs (not sure, but wondering)?

Anyway, as I was not sure what this "die size" point was about, could you further elaborate? Is that about cost? Anything else in mind?
 

Offline Electro FanTopic starter

  • Super Contributor
  • ***
  • Posts: 3320
Re: Some thoughts on building a video gaming computer
« Reply #2 on: July 15, 2020, 08:30:46 pm »
8.   For some reason (probably economics and inertia), Intel has been somewhat slow to reduce die size; conversely AMD has been more rapid to market with smaller die sizes.

How so? Can you be more explicit on that?

Intel has usually been one step forward using the finer process nodes. If die size is not reduced significantly , it's because it's influenced obviously by the content, but also the number of pads (which is pretty large on a modern CPU.) What takes area as far as "content" goes are cores of course, but also cache memory which takes a lot of area. Reducing die size would basically mean reducing the number of cores per die, and/or cache memory. Or significantly simplifying the architecture, but that would probably be a major overhaul. What exactly do you see in that for the end-user?

I've seen recent AMD CPUs with many cores are actually made of several interconnected dies (basically a 32-core CPU is made of 4 8-core dies), is that what you mean? But what's the benefit for the end-user? And does Intel not do the same with their >= 16 core CPUs (not sure, but wondering)?

Anyway, as I was not sure what this "die size" point was about, could you further elaborate? Is that about cost? Anything else in mind?

As an end user I'm not overly concerned with die size but it does seem that Intel has pushed back their plans for reaching 7nm.

Just seems like Intel has been sitting on it's laurels a bit.  I'm still a big Intel fan but it would be nice to see them move along a tad more swiftly.  Of course there is no reason for them to move any faster than they want unless they sense they have some competition.

I've always used/selected Intel processors but this time I went with AMD.  I would have used Intel again (I really don't care about die size all that much) but I wanted PCIe 4.0 for the bus and specifically for the NVMe. 

Regarding die size this is fairly recent:

https://www.digitaltrends.com/computing/inel-delays-7nm-die-size-shrink-to-2021/

https://www.tomshardware.com/news/intel-roadmap-no-high-end-desktop-processors-this-year

The Cascade Lake-X family currently represents Intel in the HEDT market, but the chipmaker has been stagnant in this segment. The 14nm processors span from 10 to 18 cores, which does seem underwhelming when compared to AMD's third-generation Ryzen Threadripper (Castle Peak) chips. The core-heavy parts start at 24 cores with the Ryzen Threadripper 3960X and push up to 64 cores in the Ryzen Threadripper 3990X

https://www.eetasia.com/intels-path-to-10nm-past-present-and-future-part-2/

Summary

Intel made big bets on its 10nm process technology and its Hyper Scaling in a bit to stay ahead of the whole semiconductor industry and produce chips that would be more competitive when compared to products from its rivals. The company had to delay HVM of its 10nm CPUs by 2.5 – 3 years from the original schedule, alter its 10nm technology for its Ice Lake products, and spend billions of dollars on expanding its 14nm output to satisfy demand for its processors in 2018 – 2019. So far, Intel’s 10nm strategy has not paid off. The company does not expect shipments of its 10nm products to crossover with shipments of its 14nm products even by the end of 2020. Furthermore, Intel admitted earlier this year that its 10nm node would not be as profitable as its predecessors.

“It is and extremely challenging business with lots of risk, which is why there are very few major players left,” said a former Intel employee. “I would agree Intel slipped up, but I would not bet against them nor would I expect them to jump ahead by huge margins in the future.”

Intel has learnt its lessons and changed its approach to process development and product design. While developing its next major node, the company makes intra-node improvements to its existing fabrication technologies to extract additional gains. Meanwhile, Intel is working to shrink development cycle of its major nodes to 2 – 2.5 years. Furthermore, the company no longer ties an upcoming architecture to an upcoming process technology, but tends to develop both separately in a bid to make its products using the most viable process technology that exists to ensure time-to-market. Last but not least, with its 10nm, Intel has learnt how to deal with extreme cases of multi-patterning and since the latter is not going anywhere, the expertise will be used in the EUVL era.

“We have gained significant learnings from our initial 10nm ramp and are applying those lessons to deliver a compelling process roadmap that better balances schedule, risk, and the introduction of new innovations,” said an Intel’s spokesperson.

Intel is a huge corporation with big plans and big customers, so it is not going to skip 10nm node in favor of a more advanced one. We are going to see many products on 10nm/10nm+/10nm++. Yet, Intel seems to pin a lot of hopes on its EUVL-based 7nm process technology.

“Looking forward to 7nm, we are returning to a more traditional density scaling factor, introducing fewer interdependent innovations at the same time, reducing design rule complexity, and leveraging a more capable lithography toolset with EUV,” an Intel official said.

Intel’s first 7nm-based product is an HPC-oriented GPU codenamed Ponte Vecchio. While based on a chiplet design relying on EMIB and Foveros technologies, the Ponte Vecchio is indeed a Big GPU meant to scale out in HPC environments. Intel used to start production of chips using leading-edge technologies from small products back in the 14nm and 10nm eras, yet, it chose a different approach with its 7nm process. Perhaps, because it just had to.

“They really cannot afford another 10nm-like screw-up,” said Brookwood. “7nm needs to be a winner. Yields always matter, even if you ignore the economics. You need to have some predictability in the process or you can’t trust the reliability of the resulting output.”


Over the last few years it seems like delays have been ongoing.

https://www.extremetech.com/computing/256127-rumors-imply-intel-pushed-10nm-cannon-lake-back-2018

https://www.anandtech.com/show/11550/the-intel-skylakex-review-core-i9-7900x-i7-7820x-and-i7-7800x-tested/6
« Last Edit: July 15, 2020, 09:11:56 pm by Electro Fan »
 

Online SiliconWizard

  • Super Contributor
  • ***
  • Posts: 15926
  • Country: fr
Re: Some thoughts on building a video gaming computer
« Reply #3 on: July 16, 2020, 02:47:04 pm »
So by "die size", you were merely talking about process nodes. That was confusing, as a given CPU can be manufactured on a finer node and still have a larger die size if it's a lot more complex. The term was unfortunate IMO. Die sizes (at least for PC CPUs) have not shrunk significantly over the years. Manufacturers have just been able to fit a heck of a lot more on one die. A "die" is just the piece of wafer for a single IC. Looks like this term of "die size" comes from some articles that simplify things, but the usual technical term would be "process node". Guess it wouldn't ring a bell for many though.

Anyway, Intel has never lagged much IMO when it comes to using finer nodes. You're probably referring to the latest generation (Ryzen) which is on 7nm, whereas the Core i9 is still on 14nm, and the relative (IMO) lag in going from 14nm to a finer node. But again, you have to consider what's in it for the consumer? Finer nodes have their own issues, such as higher leakage for instance, so power consumption can become an issue. They tend to have lower yields too. There are technical issues to solve beyond mere economics.

Now for the consumer, I guess what matters is performance, cost and (for those who care), power consumption. Intel has always been a bit ahead on overall performance and power consumption. Not cost obviously. Now it looks like the new Ryzen generation changes things, I don't know by how much though. I've seen some benchmarks, but it's not that clear cut whether the high end Ryzen CPUs are really performing better than the high end Intel CPUs. Depends a lot on the benchmark. AMD CPUs are still cheaper though, but the new high-end Ryzen CPUs are more expensive than the typical AMD CPUs used to be.

As far as strategy goes, sure Intel is a big corporation with everything that comes with that. One point that matters here too is that AMD's foundry is Globalfoundries, which is independent from AMD. AMD decided to go fabless over ten years ago, which probably helped them here.
 

Offline Electro FanTopic starter

  • Super Contributor
  • ***
  • Posts: 3320
Re: Some thoughts on building a video gaming computer
« Reply #4 on: July 16, 2020, 07:43:40 pm »
Regarding semiconductor and especially microprocessor design and manufacturing - it comes with a lot of intricate complexity. FWIW, here is an excerpt from Wikipedia (maybe not the world's authority on microprocessors but I think it's a reasonable attempt at describing die size):

Die shrinks are the key to improving price/performance at semiconductor companies such as Samsung, Intel, TSMC, and SK Hynix, and fabless manufacturers such as AMD (including the former ATI), NVIDIA and MediaTek.

Intel, in particular, formerly focused on leveraging die shrinks to improve product performance at a regular cadence through its Tick-Tock model. In this business model, every new microarchitecture (tock) is followed by a die shrink (tick) to improve performance with the same microarchitecture.[2]

Die shrinks are beneficial to end-users as shrinking a die reduces the current used by each transistor switching on or off in semiconductor devices while maintaining the same clock frequency of a chip, making a product with less power consumption (and thus less heat production), increased clock rate headroom, and lower prices.[2] Since the cost to fabricate a 200-mm or 300-mm silicon wafer is proportional to the number of fabrication steps, and not proportional to the number of chips on the wafer, die shrinks cram more chips onto each wafer, resulting in lowered manufacturing costs per chip.


Whatever the definition or significance of die size might be at the end of the day I think customers are looking at several decision-making criteria including:

1. Performance - this might start with clock speed but for some users it will get to performance per core, number of threads, etc.
2. Power consumption - this might influence some users but maybe not in terms of concern about a few more Watts as much as more heat.  (Here we are speaking of CPUs; for GPUs power consumption is a much bigger consideration.)
3. Reliability and longevity - related to power consumption
4. Compatibility - I think for a long time this was something that gave Intel a huge advantage; people were confident that if anything was Microsoft compatible it had to be Intel.  Over time AMD has maybe earned itself a reasonably close second place.
5. Security - Intel has had a few hiccups but they have shown a willingness and some ability to respond with current product patches, and to apply lessons learned to subsequent designs.  Not sure how far along AMD is in this area.
6. Price - I think Intel and AMD now are close enough on the above that while customers will pay a premium for one or the other depending on their criteria and affinity, the two are now somewhat locked into a reasonably close price proximity with one another.
7. Marketing/brand awareness - at the enterprise level no one ever got fired for going with Intel Inside; at the consumer level Intel Inside is still pretty catchy but maybe AMD has earned a foothold or better.

In some (many?) cases Intel wins on price and performance - but for various reasons AMD seems to have earned some reasons to exist - possibly AMD is just more nimble and able to focus on selective market segments where it is able to press it's advantages.

Back to die size.  From my perspective, the competition in die size - whether in terms of real technical merit or simply marketing - is interesting but not likely to be the driving force in my decision regarding which CPU to purchase.  All other things being even, and maybe 10-15% more expensive for Intel, I'd go with Intel.  Certainly, Intel at less than the price of AMD is an attractive proposition (see attachment below).  Until this recent computer build, I'd never strayed from Intel and price difference was never really even a consideration.  In fact, if everything else was somehow even but the die size was 2x or maybe 3x larger for Intel vs AMD, I'd likely still go with Intel.  The reality however is that at some point die size will probably impact functionality, power requirements, and clock frequency.  But that's not what cost Intel one processor unit (I'm sure they are not crying in their beer over losing the sale of one unit).

So, we're a little off the original thread but to be clear, I just threw in die size as a passing reference in the list of things to think about when building a video gaming computer (especially for anyone asking "can AMD really compete?").  In terms of priority die size as a metric by itself is likely last on the list (except for the fact that it might eventually impact other items on the list).

What ultimately drove me to experiment with AMD was the desire for PCIe 4.0.

If I was an Intel marketing manager I'd be more concerned with getting from PCIe 3.0 to 4.0 than I would be with getting from 14nm to 10nm to 7nm.  Now if I was the engineering manager I might care more about die size, but as the consumer I like double the bandwidth on the bus more than I like half the die size.

If anyone has ever tried to back up disks running at 250 MBytes per second, I think they are going to really like 5 GBytes per second, and pretty soon even faster.  This is not all attributable to PCIe 4.0 - it's mostly NVMe - but why not build a computer that can double the bus bandwidth for roughly the same price as half the bus bandwidth? 

(And BTW, PCIe 5.0 is coming, it's only a matter of time.  PCIe 5.0 will double the bandwidth over PCIe 4.0.  Why should we care so much about doubling?  After all it's more or less Moore's Law and it's been going on for a long time - and some people think it might be nearing it's end... or maybe not.  Either way, the point is that when you double from 1 to 2 you get an impressive thing (a double), and likewise when you double from 2 to 4 you get another impressive thing (another double), but at this point you are still at 4, and only 3 up from where you started.  When you double from say 64 to 128 you are up 64 with one jump; and when you double from 128 to 256 you are up 128 in one jump.  For anyone who likes performance increases, doubling might be more cool later on the curve than earlier on the curve, I think.  In any event, each time we take a bottleneck out of the overall system we create the opportunity to go address another bottleneck or limiting factor somewhere else.  Personally, I think the jump from spinning disks to SSDs (and NVMe sitting on PCIe buses rather than SSDs connected with SATA in particular) is very cool; kind of like moving from diskettes to hard disks, but further up the curve.)

We can talk more about die size, microprocessor design, and semiconductor manufacturing (or maybe open another thread on that); back to video game computer building....

EF
 

Offline Halcyon

  • Global Moderator
  • *****
  • Posts: 6166
  • Country: au
Re: Some thoughts on building a video gaming computer
« Reply #5 on: July 21, 2020, 10:15:28 am »
My experience doesn't come from gaming but from video production, nevertheless, this whole "three-digit FPS makes you a better gamer" is completely horse shit and is the latest marketing tactic that GPU and monitor manufacturers seem to be focusing on and consumers are buying it up.

To put it simply, there's a limit to how many frames a human eye is able to detect and more importantly, how quickly your brain and muscles can react to changes in those frames.

Don't get me wrong, from 30 frames per second to 60 frames per second is a huge visual improvement and it makes for a much smoother experience. Once you start getting into that 120+ territory, it starts to make little, if any, difference. And this is assuming that your entire system from the software itself all the way through to your monitor is even capable of displaying those types of refresh rates natively.

Beyond a certain point, fast frame rates don't make you a better gamer. Practice does.
 

Offline Electro FanTopic starter

  • Super Contributor
  • ***
  • Posts: 3320
Re: Some thoughts on building a video gaming computer
« Reply #6 on: July 23, 2020, 08:01:06 pm »
My experience doesn't come from gaming but from video production, nevertheless, this whole "three-digit FPS makes you a better gamer" is completely horse shit and is the latest marketing tactic that GPU and monitor manufacturers seem to be focusing on and consumers are buying it up.

To put it simply, there's a limit to how many frames a human eye is able to detect and more importantly, how quickly your brain and muscles can react to changes in those frames.

Don't get me wrong, from 30 frames per second to 60 frames per second is a huge visual improvement and it makes for a much smoother experience. Once you start getting into that 120+ territory, it starts to make little, if any, difference. And this is assuming that your entire system from the software itself all the way through to your monitor is even capable of displaying those types of refresh rates natively.

Beyond a certain point, fast frame rates don't make you a better gamer. Practice does.

That's more or less the conclusion of the video toward the top of the original post.
 

Offline Electro FanTopic starter

  • Super Contributor
  • ***
  • Posts: 3320
Re: Some thoughts on building a video gaming computer
« Reply #7 on: July 24, 2020, 05:57:34 pm »
Looks like the market thought it might be better to get to 7nm sooner than later.

https://markets.businessinsider.com/news/stocks/intel-stock-price-plunges-next-generation-chip-release-delay-earnings-2020-7-1029428750

"We are seeing an approximate six-month shift in our 7nm-based CPU product timing relative to prior expectations," Intel CEO Bob Swan said in a statement. "The primary driver is the yield of our 7nm process, which based on recent data, is now trending approximately twelve months behind our internal target. "

He continued: "We have identified a defect mode in our 7nm process that resulted in yield degradation. We've root-caused the issue and believe there are no fundamental roadblocks, but we have also invested in contingency plans to hedge against further schedule uncertainty."

https://hothardware.com/news/intel-7nm-cpus-late-2021-90-percent-market-share-days-over

In one day Intel's valuation goes down ~15% and AMD's goes up ~16%.

I'm still a big fan of Intel and I think the CEO is doing the right thing to better set expectations on schedules and what to emphasize and prioritize.  The good news for both companies is that the market for CPUs and other semiconductor products isn't going anywhere but up for a long time. 
 

Offline LeonR

  • Regular Contributor
  • *
  • Posts: 160
  • Country: br
  • PC hardware enthusiast
Re: Some thoughts on building a video gaming computer
« Reply #8 on: August 04, 2020, 01:30:16 am »
2.   The FPS you are going to achieve will depend on the refresh rate of the monitor, which in turn will be heavily dependent on the resolution.  The impact of FPS (and some related specs) is the subject of this youtube video:


Unless you're using v-sync or a hardware sync like G-Sync (nVidia) or FreeSync (AMD/nVidia), game engine frame-per-second (FPS) rate isn't linked to monitor refresh rate (frequency). You can have a gaming running at 300+ FPS while using a basic 60Hz monitor. Also most action-intensive games like first-person shooters (FPS genre) benefit from running at higher FPS.

Frame-per-second isn't the only performance metric for games - Frame Times (measured in miliseconds) are also important. Some systems (or game engines) may behave erratically and take longer to render a frame by whatever reason, leading to a delay and increased frame time that is known as an stutter. Having any sort of constant stuttering detracts greatly from the gaming experience. Lots of gamers have trouble with this.

Under ideal circumstances, frametime is tied to FPS. 60 FPS = ~16.6ms frametime. An example of a erratic performance on this was Insurgency Sandstorm: although RivaTuner (a real-time, on-screen performance metrics display) reported somewhat good FPS numbers (90+), game always had 30ms+ frametimes, resulting in very poor overall gaming performance. Dev fixed this some time ago, though.

(haven't watched the video, btw)

4.   In addition to having a monitor that can achieve 240 FPS (by supporting 240 Hz or better refresh rate), a/the key link in the chain is the GPU.  It is possible that a very high-end GPU (Nvidia 2080 Super, etc.) can output more than 240 FPS but the monitor will then be the liming factor.  Some monitors can support 280 Hz and therefore 280 FPS.  In some cases you can see even higher (than 280) FPS numbers displayed as a performance readout on the monitor but this represents what the GPU is outputting vs. what the monitor is doing.

Just as I've stated before, unless you're using v-sync or a hardware sync, game FPS and monitor refresh rate aren't tied.

7.   Intel CPUs are holding their own in terms of performance when using relatively few cores and few threads but as application software is able to use relatively more cores and threads the AMD processors become increasingly competitive.  How much the number of cores and threads supported will impact game performance depends on how the game (or other application) was written.

AMD Ryzen actually surpassed Intel Core i-series IPC with Ryzen 3000. There are benchmarks on tech sites with. Intel got lazy by selling the almost same 4C/8T chip for many generations (2nd to 7th Core i) and got

9.   One design aspect that AMD has a clear current (but probably temporary) lead is in the ability to support PCIe 4.0.  A PCIe 4.0 bus will deliver twice the bandwidth of a PCIe 3.0 bus.  One area where this is starting to become an advantage for AMD (and AMD motherboard makers, such as Asus and others) is in system storage.  NVME drives that support PCIe 4.0 are now available.  They can achieve sequential reads and writes of approximately 5 GBytes per second – which works out to be about 4 minutes to write a TeraByte.  This is somewhat gated by the controller chip functionality that appears to be ready for another step forward (maybe this summer) which will enable PCIe 4.0 NVME drives to achieve about 7 GBytes per second.  For games this is probably not a big deal as the game gets loaded from the drive into RAM and then the GPU becomes the primary workhorse – but if you want to make a screaming fast general purpose computer, NVME and PCIe 4.0 are the way to fly.  As a frame of reference, a Western Digital Black 7200 RPM SATA drive might achieve on the order of 270 MBytes per second (not even close to top of the line NVME).  Overall, depending on whether the data is being read/written sequentially or randomly and in what data sizes, NVME drives will perform 10-100 times faster than conventional electro-mechanical drives.  NVME might not provide the same very long-term data durability of conventional drives but for many/most users it will probably last as long as the initial computer it is supporting.

There's a good reading at Puget for this:

https://www.pugetsystems.com/labs/articles/Does-PCIe-Gen4-improve-GPU-performance-in-video-editing-apps-1565/

About disks: SATA SSDs still hold their own against NVMe models for daily usage, considering overall performance and game loading times. Although M.2 NVMe drives have some nice performance numbers when compared to SATA counterparts, they don't scale up most of the time.



10.   Before selecting your RAM, check with the motherboard manufacturer’s Qualified Vendor List to see which specific makes and models of RAM have been tested for compatibility with the motherboard.

It may be worth taking a look, but considering the thousands of combinations and new RAM modules getting released only a very small fraction of them are on those lists. Most of the time the RAM you bought will work without any issues (considering the model you chose have specs that are compatible with your motherboard).
 

Offline free_electron

  • Super Contributor
  • ***
  • Posts: 8550
  • Country: us
    • SiliconValleyGarage
Re: Some thoughts on building a video gaming computer
« Reply #9 on: August 04, 2020, 03:22:04 am »
something 6502 based. all the best games were written for 6502's ...
Professional Electron Wrangler.
Any comments, or points of view expressed, are my own and not endorsed , induced or compensated by my employer(s).
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf