Then you may not have looked close enough. I did test my examples before posting them. And in the case of Lightroom the files I tested on resided in RAM cache and most of my RAM was still unused. Switch back and forth between Gallery/Evaluation (E) and Develop (D) and watch how your machine lags behind.
I am not saying that Lightroom and Photoshop cannot be well worked with on modern PCs. I am saying that for many tasks throwing more cores at them increases performance less than throwing more IPC/frequency at them. I wish it was different, because that would literally just mean to throw money at them for better performance instead of needing new tech.
I am not noticing this after a few 100 RAWs after the last airshow I went to. I have tried SB Core i5,IB Core i7 and Haswell Core i7 chips,and honestly much of a sameness. The biggest upgrades in performance for me in day to day usage were the upgrades to the disk systems.
32-bit is a problem there. But whenever my Firefox starts lagging, it's usually a matter of CPU load (1-2 cores out of 4) than RAM. The instance of Firefox I am currently typing in uses less than 1 gb of RAM, even with lots of tabs open. CTRL-Tabbing through the tabs still is not instant. Of course I am more sensible to lag, because I professionally work with lag (or the lack thereof).
Neither am I but the lag is more to the way Firefox handles RAM - anytime it lags look at the RAM usage,and memory leaks on Firefox especially with plugins installed has been a known on and off issue for yonks - CPU usage is not really as big a problem. Chrome in comparison has tended to be better upto a degree IMHO.
Again, all of the applications I listed usually are not limited by my I/O or RAM. When I have to wait for them it's mostly because of CPU and then often only 1 or 2 of my 4 cores are fully loaded.
Which again I am not seeing - not to the extend that a SB or Haswell CPU made much of any difference IMHO.
Put a folder with over 10k small files into your GDrive folder and watch how it hogs a single CPU at full load for minutes before any internet transfer even begins to happen.
Again the transfer is limited by disk I/O and speed of your internet connection and I should know having 100+ GB of online backup for pictures and stuff using such services. The CPU is not making anywhere as much difference - not to the extend that a SB or Haswell CPU made much of any difference IMHO,and in terms of the whole length of the process especially from a new PC.
You also need to consider,Cloud services are also as limited on the providers side.
What do you think they are using for storage - most likely to be HDDs still and those often rated for reliability and not throughput.
What CPUs are they using??
Google is moving towards using more ARM based solutions. Now ,it might be not for the servers behind the storage side of the business,but it does indicate companies are more worried about running costs than absolute performance.
Even,online games which are touted as the bastion of the importance of single threaded performance,can be massively limited by server side performance.
Games like Diablo 3 and Planetside 2 all run better on Intel chips,but during the massive firefights in both games,performance will just plummet even with a 70Mbps UP/ 20Mbps DOWN connection and that's everyone from people with a Core i7 4790K to an FX6300 I know,and in many cases is more the server side.
You only have to look at EVE:Online and time dilation during huge battles.
The question was about desktop performance. Once an SSD and much RAM are used the biggest bottleneck is the CPU. And then it more often then not is capped by the use of only 1-2 cores out of whatever number you throw at it.
But not to the extent it makes such a massive difference going between CPUs being released within two generations of each other. Its one thing going from an Atom to a Core i7 6700K,but another thing going from maybe a Core i5 2500K to the same CPU.
Even then after doing more budget builds or upgrading laptops with a faster drive like an SSD,for most general purpose stuff people do,I found them much of sameness.
If anything some of the biggest gains we have seen moving to newer CPUs is more modern extensive support,which can actually help massively in certain niche task.
I have worked with very CPU intensive stuff at uni,and we are talking MASSIVE data sets,and things that can take hours,days or weeks to process. Some of the image deconvolution stuff was the destroyer of PCs.
Add the use of things like Photoshop,PS software and even custom stuff we needed to use(and friends too),often running on SLOWER PCs than what a lot of enthusiasts have,it makes me wonder how we managed to get to publication deadlines on time!! :thumbsup:
Remember,how popular tablets became with the iPad?? Yet,many of these tablets had worse single threaded and multi-threaded performance than an Atom and plenty of laptops still destroy them for pure CPU processing speed.
Yet,people like Apple worked on reducing other bottlenecks in the chain,and using IGPs to offload certain tasks more effectively which meant for the average Joe or Jane it felt more responsive than a netbook or bog standard laptop.
I think we might have to agree to disagree otherwise we will go round in circles,but IMHO there are more factors which need to be taken into consideration and people focus way too much on CPUs only.
Facebook
Twitter