Gadgetory

welcome back to our unboxed today's video has been heavily requested so I've made it a priority and it's the first of these second generalized and exploration videos that I have planned since the release of the second gen rising cpus many of you have been asking me for an IPC comparison with intel's coffee like cpus for those of you unaware IP C stands for instructions per cycle and it is a good indicator of a CPUs performance the coffee like CPUs offer high IPC coupled with high operating frequency and that really is the best combination for maximum performance although AMD is clearly trailing when it comes to frequency they appear to have really closed up on intel's IPC performance and that's likely why many of you have been requesting this test before we get into the IPC comparison this video has been sponsored by Squarespace whether you need a domain website or online store make it with Squarespace for more information please check the link in the video description to see just how much headway AMD's made here we're going to neutralize as many variables as we can while also keeping things as realistic as possible the first and most obvious step is to remove call frequency from the equation and to do this we've locked all CPU cores at 4 gigahertz any type of boost technology has been disabled the course cannot go past 4 gigahertz and all cores are again locked at 4 gigahertz the second generalising CPUs have been tested on the asrock x4 70 Taichi ultimate and the coffee-like CPUs on the asrock said 370 Taiji both configurations used the same g.skill flex ddr4 3,200 memory using the extreme memory profile and the same MSI GTX 10 a DTI gaming extra was used for all the testing this is purely for a science type benchmark and is in no way buying advice the coffee-like cpus do have a clock speed advantage there's simply no getting away from that fact and for real-world performance please refer to our recent risin 5 2,600 2,600 X and 2,700 X reviews for this test I've included results for the Intel Core i7 87 okay 8600 k horizon 7 2700 X 2600 X and Verizon 7 1800 X along with the 1600 now the 1600 X 2600 X and 8700 K all have the same CPU resources six cores with 12 threads the 1800 X and 2700 X have an advantage being they are eight core 16 thread CPUs well the 8600 k is at a disadvantage because it's a six core six threaded CPU so please keep all that in mind as we proceed right so let's get to the results starting with the sustained memory bandwidth test here we see that the first and second generation C abused are very similar with a bandwidth of around 39 gigabytes per second meanwhile using the exact same memory the coffee-like CPUs are limited to around 33 gigabytes per second and this is a 15% reduction in bandwidth when compared to the rise in CPUs moving to the Cinebench r15 results we see at the 2600 X scores 4 percent higher than the 1600 for the multi-threaded test and 3 percent higher for the single thread test then as we look at the 8700 K we see that it is 4 percent faster than the 2600 X for the single threaded test but 4 percent slower for the multi-threaded test as you might have expected clock for clock the 8 core 6 don't read rise and cpus easily beat the multi-threaded score of 87 are ok I've included them simply because I had the results depending on demand I could update this test with the core i7 78 20x for example next up we have the PC mark 10 video editing scores and this is a more lightly threaded test though we did previously see a noteworthy difference between the 1600 X and 1800 X as a result we see a solid 10 percent jump from the 16 hundreds to the 2600 X and this place is AMD on par with Intel in terms of IPC performance for this test like what we saw with these Cinebench r15 benchmark when maxed out SMT appears to be more efficient than Intel's HT technology here the 6100 X was fast nearly 700 K by a three and a half percent margin while the 2600 X was eight percent faster and that really is a noteworthy margin right there next that we have Excel and here the 87 her cake was around 3% quick up than the 1600 X when matched clock for clock however the 2600 X is able to match the 87 okay with the same completion time of 2.8 five seconds and pretty impressive the handbrake test we run doesn't leverage the AMD rising CPUs that well and here we see at the 2600 X is only able to match the 80 600 K and therefore is 15% slower than the 80s 700 K moving on to the corona benchmark we set the 2600 X is able to reduce the render time by 8% when compared to the 1600 X and this MANET was just 3 percent slower than the 8700 K so Intel still holds an IPC advantage in this test but it is very minor next up we have blender and here the 2600 X was just two and a half percent quicker than these 1600 X and this made it 4 percent slower than the 80s of her ok not a huge difference and again while Intel still holds an IPC advantage in this test it is less than 5% now using the v-ray benchmark we set the 2600 X improves on the 1600 by 4% margin and that met it was just a single percent slower than the 87 okay so basically on par here ok so time for some gaming results and this is where things come unstuck for AMD as we've talked about in the past Intel's low latency ring bus is just better suited for gaming and we see this when comparing their own mesh interconnect architecture designed for the higher core count CPUs the Infinity fabric suffers the same problem as the mesh architecture and it's not until gaming CPUs require way more cause this problem will go away for AMD so while the 2600 X does improve on the 1600 X by an 8 percent margin and ashes of the singularity it's still a whopping 11% slow of nearly 700 K a couple that with the fact that Intel CPU is clocked much higher this can blow that margin out to over 20% at times moving on to Assassin's Creed origins and here we see a mere 2% increase for the 2600 X over the 1600 X while the 87 or ok is a further 14 percent faster the margin is slightly reduced with the higher quality preset but still the 87 Andrew K is 12% fast in the 2600 X when comparing the average framerate when testing with battlefield 1 using the ultra quality preset we see the 2600 X is 9% fast and the 1600 X but still 7% slower than the 87 ok the margin is completely blown out of the water with the medium quality preset as the gtx 1080i is now able better stretch its legs here the 2600 X again off in a 9% increase over the 1600 X but is now 10% slower than 87 okay which still appears to be GPU limited it's a similar story when testing with far cry 5 the 2600 X is again 10% faster than the 1600 X which is a huge improvement but even so it's still 8% slower than the 8700 K moving on to power consumption and firstly just let me quickly point out that this isn't the most realistic test and many of the power are saving features are disabled in our 4 gigahertz clock for clock comparison it's also not the most scientific because I have had to increase the voltages for the rise in CPUs above spec to stabilize and all core overclock of 4 gigahertz so having said that we see at least 1600 X and 2600 X systems consume the exact same amount of power meanwhile the 8700 K system consumed 3 percent less power making it slightly more efficient under these test conditions power consumption when testing with farcry was much the same across the board all CPUs push total system consumption to around 380 watts here we see a 10% reduction in power consumption we're moving from the 1600 X to the 2600 X in our blender workload that is an impressive improvement for the 2600 X but even so it still consumes 21% more power than the idea seminar ok then finally this time when testing with handbrake we set the 2600 X actually pushed total system consumption 7 percent higher than that of 1600 X and a whopping 32% higher than the 8700 K okay so we just saw a number of interesting results as we saw in the day 1 reviews despite a rather large clock speed deficit the second generalize and CPUs aren't often that far behind there Intel rivals in application benchmarks and here we see why that is when comparing them clock the clock at 4 gigahertz in application benchmarks such as Cinebench r15 we see that the single core performance is down just 3% but where SMT is well leveraged aim D can be as much as 4% faster we found that AMD was 3% slower in our corona benchmark and much the same in our Excel IV rate and video editing tests then while it was 15% slower and our handbrake test it was also 8% faster for the PC market and physics test of course there is still the matter of gaming and I bet a few AMD fans were hoping we could put most of the gaming performance deficit down to clock speed sadly though that isn't the case one issue here is how AMD connects its cause or rather for CCX modules intel's ring bus is a low latency method that always sees resources access to bio the shortest path however as you add more cause the Rings grow in size and more rings are required to connect all the cores and this sees efficiency go out the window there's more to it but basically Intel needed a method for connecting large amounts of course like 28 of them for example at this core count the mission to connect architecture is superior however four six eight and even ten core CPUs we already know the mesh interconnect is an inferior solution and this is why the core i7 is 7800 X 78 20x + 17 900 X all gets smoked by the 8700 K when it comes to gaming this is because the 8700 K has an average latency for about 40 nano seconds between cores whereas the 700 X is more like 70 to 80 nanoseconds the rise in CPUs are a little more complex as the cores within the CCX module have a similar core to call latency to that of the 8700 ok and this is regardless of the ddr4 memory speed however once you exit the CCX module the cortical latency increases to around a hundred and ten nanoseconds and that's with ddr4 30 to 200 memory the CCX to c6 core latency is reduced with faster memories aimed is infinity fabric is tied to the memory clock rate and lower latency DRAM also helps immensely another issue is the games themselves almost all games are designed to work on just a few cores and only now are we starting to see some effort made to try and break tasks up into pieces so they can be run in parallel obviously prior to aim DS rise and release games were designed and well almost exclusively optimized for Intel CPUs and that meant working with quad cores that is slowly starting to change now so rising gaming performance will improve with time just don't expect it to be wiping the floor with ring bus CPUs any time soon anyway in terms of IPC performance amy has certainly closed up the gap and the improved cache latency has also really helped there certainly a number of benefits to buying a second generalize and CPU over a coffee like CPU so it is going to be exciting to see how the battle unfolds in 2018 and beyond there's also a number of benefits to squarespace's all in one platform it allows you to quickly and easily create a beautiful hassle-free website there's nothing to install patch or upgrade ever you don't need to know a thing about stuff like PHP CSS HTML they're all just acronyms for things you don't need to worry about if you build with Squarespace they're beautiful designer templates make creating a powerful online identity easy each template is a starting point for a wide range of projects whether you're pursuing a side hustle promoting your main gig Squarespace also provides award-winning 24/7 customer support go to squarespace.com forward slash Harper unboxed single word no space and start your free trial today and receive 10% off your first purchase of a website or domain and that is going to do it for this one if you did enjoy the video then be sure to hit the like button for us appreciate that subscribe for more content and if you appreciate the work we do have our box then it can see that supporting us on patreon thanks for watching I'm your host Steve and I will see you again next time