WEBVTT 00:00:00.000 --> 00:00:15.880 Well, Lylok, you've done it again. I don't think the PCB will be troubling us any time soon. 00:00:15.880 --> 00:00:21.220 Quite indeedily, Pluton. Quite indeedily. But just tell me, how did you manage to crack the case? 00:00:21.220 --> 00:00:27.080 Ms. Kateson's got her shawl back, and you've managed to determine which of our pre-built gaming computers is the best. 00:00:27.080 --> 00:00:37.240 Well, it was quite elementary. You see, it was near the end when Scotland Yard examined our pre-configured computing machines. 00:00:37.240 --> 00:00:44.120 Adele, we meet again. And this time you're very similar to last year. 00:00:44.120 --> 00:00:51.000 The XPS8960 still requires a screwdriver to access the side panel, but is otherwise mostly tool-less. 00:00:51.000 --> 00:00:58.920 You get two fans, front and rear, with space for a third, along with decent cable management, including pre-run cables for adding a hard drive, 00:00:58.920 --> 00:01:02.960 as well as a GPU brace that could probably survive an earthquake. 00:01:02.960 --> 00:01:10.400 The proprietary Z690 motherboard will cause some headaches if you want to upgrade the power supply later, which you may very well want to do, 00:01:10.400 --> 00:01:14.880 but is well-featured with two M.2 slots and on-board Wi-Fi. 00:01:14.880 --> 00:01:20.200 The only major quality complaint that I have here is the cheapo flower-type CPU cooler. 00:01:20.200 --> 00:01:28.560 There's a Core i7-14700 non-K under this thing, and that would live a happier, cooler life with even a basic tower heatsink, 00:01:28.560 --> 00:01:35.120 like, oh, I don't know, the one that they give you if you ordered the K variant of that same CPU. 00:01:35.120 --> 00:01:44.640 Otherwise, the specs are about what we would expect at this price, except for the 460-watt bronze efficiency power supply. 00:01:44.640 --> 00:01:53.840 At least Dell's system arrived fully intact, though. I don't know what happened to our iBuy power system and shipping, but aside from the thumb screw that was lodged in the I-O, 00:01:53.840 --> 00:02:02.880 upon closer inspection, we found that nearly all of our motherboard screws were loose, and two of the ones that weren't were cross-threaded. 00:02:02.880 --> 00:02:11.720 As for the case, it's not particularly tool-less, but it's also not difficult to work in, and we got four 120mm ARGB fans, 00:02:11.800 --> 00:02:20.720 though they use a proprietary connector, which is a bit of a downer, and for some reason, we're not configured to sync with our motherboard, or even to each other. 00:02:20.720 --> 00:02:32.000 Cable management is decent, with nothing loose or hanging around, and this time, we got a much better CPU heatsink with what looks like an iBuy power badge to Hyper 212 ARGB tower cooler. 00:02:32.000 --> 00:02:36.280 You guys were so close, you just had to screw down the motherboard. 00:02:36.280 --> 00:02:42.080 Oh well, as for specs, we got a strong CPU, dual-channel memory, and ample storage. 00:02:42.080 --> 00:02:53.200 Too bad there was a little mix-up with iBuy power's unauthorized reseller on Amazon, and we didn't get the 4060 Ti that was supposed to be in this system. 00:02:53.200 --> 00:03:04.560 Moving on to main gear, the MG1 case is largely unchanged since our previous secret shopping spree, and if tis not broken then, nay, it need not be fixed. 00:03:04.560 --> 00:03:09.600 With that said, a filter behind the removable faceplate would be a nice touch. 00:03:09.600 --> 00:03:17.880 Cable management was a strong point, though, with no shortage of zip ties, though that might be a pain point when you're upgrading later and you have to start cutting them off, 00:03:17.880 --> 00:03:29.160 and we got a beefy RGB fan controller requiring two SATA power connectors that supports six fans and up to eight ARGB headers for extra lighting strips or effects. 00:03:29.160 --> 00:03:32.400 It also appears our case power switch is routed to it as well. 00:03:32.400 --> 00:03:36.720 Curious, another odd choice was the anti-seg GPU bracket. 00:03:36.720 --> 00:03:45.040 It was a nice touch to have it in there, but the way it's mounted made replacing the graphics card in the system a bit of a pain, so be prepared. 00:03:45.040 --> 00:03:51.040 Main gear is our first competitor to go AMD for the CPU, which we all know is good for gaming, 00:03:51.040 --> 00:03:58.520 but the chip that they chose has literally one third as many processing cores as the last two systems we looked at, 00:03:58.520 --> 00:04:07.680 and while main gear chose some of the highest clocked memory in this round, the high-rated latency means they might not get as much performance benefit from it. 00:04:07.680 --> 00:04:14.040 As for the rest of the specs, the main standout was main gear selection of an SSD with AD RAM cache. 00:04:14.040 --> 00:04:18.080 Generally, these drives perform better and last longer. 00:04:18.080 --> 00:04:28.240 As for cooling, we got four 120mm fans and what looks like a Hyper 212 Black Edition, so you probably won't need to change any of that out anytime soon. 00:04:28.280 --> 00:04:33.560 Speaking of change, I'm happy to report that StarForge has gone through quite a bit of change. 00:04:33.560 --> 00:04:40.720 Gone is the deep-cooled CG650 from last time, and in its place, we have the beautiful Phanteks XT View, 00:04:40.720 --> 00:04:45.000 a gorgeous and mostly tool-less case that's easy to work in. 00:04:45.000 --> 00:04:53.440 By the way, those of you who suspected that StarForge's fans were all set to exhaust were hoodwinked by their use of reverse-blade intakes. 00:04:53.440 --> 00:05:02.600 There is an additional 120mm fan that is exhausting out the rear, but the rest of them are all daisy-changed together for some positive pressure goodness. 00:05:02.600 --> 00:05:10.400 Curiously, this case does include an RGB controller, but StarForge opted to plug into the motherboard for control instead. 00:05:10.400 --> 00:05:14.640 Another thing they didn't use was the case's included mounting hardware. 00:05:14.640 --> 00:05:21.080 They clearly used something to assemble the computer, but we were left with the full set of screws and zip ties. 00:05:21.120 --> 00:05:24.760 Not that we'll need them since the cable management was solid. 00:05:24.760 --> 00:05:30.000 I'm still not a fan of this uncapped ARGB cable that could short out on something, 00:05:30.000 --> 00:05:34.840 and some of these SATA cables could be tucked away a little nicer, but it's not bad at all. 00:05:34.840 --> 00:05:42.040 We need to talk about the CPU cooler, though. The contact frame that's included with their Arctic cooler is a nice touch, 00:05:42.040 --> 00:05:47.440 but StarForge, you might want to let your techs know that you don't need the full tube of thermal paste, 00:05:47.440 --> 00:05:53.360 and ours, while secure enough, was actually not torqued down all the way. 00:05:53.360 --> 00:05:58.080 As for the heatsink, some of you noticed that it was a skew after I ripped out the packing foam, 00:05:58.080 --> 00:06:04.000 but I'm happy to report that while it looks a tad wonky, it didn't seem to affect performance. 00:06:04.000 --> 00:06:11.760 Under our cooler, we find our most pedestrian CPU yet mounted to the most bulk of motherboards we've seen yet, 00:06:11.760 --> 00:06:16.200 along with some other kind of mediocre stuff. Except for the GPU. 00:06:16.200 --> 00:06:23.720 Our only Radeon card this round is an RX6650XT8 gig from three years ago. 00:06:23.720 --> 00:06:29.160 We'll also be exploring the performance of this system with the 7700XT upgrade 00:06:29.160 --> 00:06:32.640 that we would have gotten for the $300 we had left in our budget, 00:06:32.640 --> 00:06:38.000 but our value calculation will be based on the system that they sold us. 00:06:38.000 --> 00:06:43.600 Next up is HP, with a new case that sports these nifty, stealthy Wi-Fi antennas 00:06:43.640 --> 00:06:50.080 and a perforated front panel for better airflow for our dual 140mm intake fans. 00:06:50.080 --> 00:06:55.640 We've got a 120mm exhaust in the rear, and all three fans are wired directly to our motherboard headers 00:06:55.640 --> 00:07:02.840 instead of into the perfectly good RGB fan hub, where they did decide to route the RGB cables for the fans. 00:07:02.840 --> 00:07:08.200 Cable management on the back isn't terrible, but on the front, come on HP. 00:07:08.200 --> 00:07:11.240 Either use a nicer case with the power supply shroud, 00:07:11.280 --> 00:07:17.520 you know, like the one you sent us last time, or just take a little bit more time to tidy this up. 00:07:17.520 --> 00:07:23.240 On the plus side, unlike Dell, HP has basically no proprietary weirdness going on 00:07:23.240 --> 00:07:26.800 that would prevent you from changing parts or upgrading in the future. 00:07:26.800 --> 00:07:35.120 And upgrade in the future, you shall. On paper, this is our weakest system yet, pretty much across the board. 00:07:35.120 --> 00:07:40.400 Okay, I guess the slightly less basic CPU cooler and the 500W power supply 00:07:40.400 --> 00:07:44.920 are technically ahead of Dell, but not by much. 00:07:44.920 --> 00:07:49.880 CyberPower's case is designed by Coolermaster and resembles a Masterbox 5 Lite, 00:07:49.880 --> 00:07:59.280 but with some better ventilation at the front. It includes four 120mm RGB fans, though surprisingly no RGB fan hub. 00:07:59.280 --> 00:08:03.480 Cable management is mediocre, no thanks to the board or the case, 00:08:03.480 --> 00:08:09.400 but at least they tried, other than this unholy chain of 4-pin Molex connectors for those fans. 00:08:10.280 --> 00:08:14.120 Where they could have tried harder is the specs, decent CPU, 00:08:14.120 --> 00:08:20.800 and surprisingly this 600W Apivia power supply didn't fail our testing, more on that later, 00:08:20.800 --> 00:08:27.560 but a single 16GB stick of memory in 2025 and this subpar flower cooler? 00:08:27.560 --> 00:08:32.480 We didn't do formal acoustic testing, but I can tell you this one gets pretty loud, 00:08:32.480 --> 00:08:39.040 almost as loud as the Dell. Wrapping up our physical evaluation, in third place we've got CyberPower, 00:08:39.040 --> 00:08:44.360 though I wouldn't mind some more amenities, especially front USB Type-C. 00:08:44.360 --> 00:08:48.240 As for the win, I've got to give it to Manegear overall. 00:08:48.240 --> 00:08:54.120 StarForge did a great job with cable management and packing in the Phanteks hardware was a nice touch, 00:08:54.120 --> 00:08:57.360 but Manegear looked ever so slightly cleaner, 00:08:57.360 --> 00:09:00.880 not to mention they didn't go overboard on the thermal compound. 00:09:00.880 --> 00:09:07.600 Of course, everything we've talked about just now, including the over application of thermal compound, is just looks. 00:09:07.600 --> 00:09:11.960 I've got a feeling that the beauty of these machines might be only skin deep 00:09:11.960 --> 00:09:16.360 and the order might change when we look at performance and value. 00:09:16.360 --> 00:09:23.440 But first, software. Well Dell, how up to date and gaming ready are you this year? 00:09:23.440 --> 00:09:29.280 According to my report, the BIOS and Windows versions were both from November of 2024, 00:09:29.280 --> 00:09:35.720 so just a few months before our ship date. However, our GPU drivers were from July. 00:09:35.720 --> 00:09:39.080 Like, of 2024! 00:09:39.080 --> 00:09:43.560 Resizable Bar, an important feature for modern GPUs, was enabled, 00:09:43.560 --> 00:09:49.760 but we found no trace of XMP memory overclocking in our antiquated looking BIOS. 00:09:49.760 --> 00:09:55.640 With that said, our memory was running at its maximum rated speed of 5600 megatransfers per second, 00:09:55.640 --> 00:09:58.800 so I guess that's fine. 00:09:58.800 --> 00:10:03.520 Our recovery environment and ability to refresh the PC were both configured properly, 00:10:03.560 --> 00:10:06.800 and while there were some Dell applications pre-installed, 00:10:06.800 --> 00:10:10.480 they were for the most part not configured to start up without permission, 00:10:10.480 --> 00:10:14.840 so we only had to remove McAfee during our de-bloating step. 00:10:14.840 --> 00:10:21.240 Let's move on to iBuyPower. For those who missed part two, this is the kind of weird PC 00:10:21.240 --> 00:10:26.640 that Miss Kateson accidentally bought from an iAuthorized third-party seller on Amazon 00:10:26.640 --> 00:10:32.800 when iBuyPower directed her to get the system from Amazon rather from them directly. 00:10:32.800 --> 00:10:36.360 And this is an odd one. Our BIOS isn't too old. 00:10:36.360 --> 00:10:41.440 It's from September of 2024, but our Windows install is ancient. 00:10:41.440 --> 00:10:46.040 I'm talking a 22H2 build from a year ago. 00:10:46.040 --> 00:10:51.120 So, I mean, it's been getting security updates since the latter half of 2022, 00:10:51.120 --> 00:10:56.240 but come on, 23H2 has been out for kind of a minute now. 00:10:56.240 --> 00:11:02.000 And our Intel management and GPU drivers are both similarly out of date. 00:11:02.000 --> 00:11:05.920 Looks like someone hasn't been updating their prep image. 00:11:05.920 --> 00:11:09.200 Thankfully, not only does the recovery environment work, 00:11:09.200 --> 00:11:16.160 an improvement over last time for iBuyPower, but there was zero bloatware and no odd folder remnants on our drive, 00:11:16.160 --> 00:11:19.640 both of which are nice to see. But wait a tick. 00:11:19.640 --> 00:11:24.440 What's this? Oh, iBuyPower, how could you? 00:11:24.440 --> 00:11:27.600 XMP is disabled out of the box 00:11:27.640 --> 00:11:33.160 and our CPU cooler tuning is set to a lowly 65W. 00:11:33.160 --> 00:11:36.600 Rebar may be on, but you guys aren't getting any points for that 00:11:36.600 --> 00:11:39.800 since this MSI board enables it by default. 00:11:39.800 --> 00:11:45.280 What a mixed bag this has been. Let's see if main gear can do any better. 00:11:45.280 --> 00:11:51.080 Considering their BIOS, Windows, and driver versions are all from 2025, 00:11:51.080 --> 00:11:55.560 most from the very same month that we ordered the machine, I'd say they have. 00:11:55.560 --> 00:11:58.800 Expo memory tuning and rebar are both enabled. 00:11:58.800 --> 00:12:04.120 There's no antivirus or additional bloat to remove and no random leftover prep folders. 00:12:04.120 --> 00:12:07.960 Magnificent. But it's not all good news. 00:12:07.960 --> 00:12:12.400 Our system had a little tiny overclock that was pre-applied. 00:12:12.400 --> 00:12:15.560 Nothing major, just a PBO boost of 100MHz 00:12:15.560 --> 00:12:19.360 and our curve optimizer set to all cores with a minus 15 offset, 00:12:19.360 --> 00:12:22.920 which sounds fine. It's just that it didn't work. 00:12:22.920 --> 00:12:28.720 We did try a few adjustments, but our system outright refused to boot with these settings enabled, 00:12:28.720 --> 00:12:34.080 so we ended up turning them off, keeping only Expo enabled for our testing. 00:12:34.080 --> 00:12:38.880 Sorry, main gear. While I do believe you that it probably worked during QC, 00:12:38.880 --> 00:12:42.640 alas, tragedy must have struck the system somewhere along the journey, 00:12:42.640 --> 00:12:47.640 which highlights the pitfalls of factory overclocking systems. 00:12:47.640 --> 00:12:52.880 What about Starforge, though? Well, it appears they've done the opposite of main gear 00:12:52.880 --> 00:12:58.960 and updated very little. Our BIOS and GPU driver versions are from late 2024 00:12:58.960 --> 00:13:03.320 and Windows hasn't been updated since late 2023. 00:13:03.320 --> 00:13:08.240 Oh my goodness, and the Intel Management Engine driver is older still. 00:13:08.240 --> 00:13:12.240 Pity. They also left a slew of leftovers on the system, 00:13:12.240 --> 00:13:16.640 like wallpapers and animations. But on the bright side, look, 00:13:16.640 --> 00:13:19.880 outdated Windows and drivers can be dealt with by the user 00:13:19.880 --> 00:13:26.000 and there is very little bloat otherwise. And XMP and Rebar were both enabled. 00:13:26.000 --> 00:13:29.840 Moving right along, let's talk about the system that almost didn't get here. 00:13:29.840 --> 00:13:37.360 HP is an odd duck this year. The BIOS version of our machine, F.08, isn't available from their website, 00:13:37.360 --> 00:13:42.840 so we're not sure when it's from, but we appear to be just one revision behind F.09, 00:13:42.840 --> 00:13:47.240 which released earlier this year, so it's probably from 2024, 00:13:47.240 --> 00:13:51.080 like our Windows version, chipset, and GPU drivers. 00:13:51.120 --> 00:13:57.320 But if you look closely at HP's website, you'll notice that we're on the latest versions of some of these. 00:13:57.320 --> 00:14:02.440 HP, are you suggesting that AMD's most recent chipset drivers, 00:14:02.440 --> 00:14:07.080 after multiple new CPU launches, are several months old? 00:14:07.080 --> 00:14:11.400 Inconceivable. No, though, actually inconceivable. 00:14:11.400 --> 00:14:17.320 What this means is that HP is not only not updating the images that they use to install their systems, 00:14:17.320 --> 00:14:21.880 but they aren't updating their own driver database and website, 00:14:21.880 --> 00:14:25.920 for shame. Also, while Rebar was enabled, 00:14:25.920 --> 00:14:30.200 our recovery environment works, and they've left no sysprep traces, 00:14:30.200 --> 00:14:35.160 they failed to enable XMP and had a bunch of unwanted software installed, 00:14:35.160 --> 00:14:39.360 like Dropbox Promotion, MyHP, Omen Audio Control, 00:14:39.360 --> 00:14:42.840 and Gaming Hub, and, huh, McAfee. 00:14:42.840 --> 00:14:47.640 We removed McAfee before testing, by the way, per our standardized process. 00:14:47.640 --> 00:14:52.040 Finally, there's CyberPower. Their BIOS is fairly recent. 00:14:52.040 --> 00:14:55.760 XMP and Rebar were both enabled, their recovery environment works, 00:14:55.760 --> 00:15:00.080 and system prep was nice and clean, which puts main gear on top, 00:15:00.080 --> 00:15:04.160 and cyber power in a clear second place when it comes to system prep, 00:15:04.160 --> 00:15:07.240 with Dell making a surprise podium appearance 00:15:07.240 --> 00:15:12.520 thanks to their non-ancient software. But, who will win the all-important 00:15:12.520 --> 00:15:15.080 gaming performance carriage race? 00:15:16.080 --> 00:15:17.080 This is it. 00:15:23.080 --> 00:15:27.080 This can't be right. Do my eyes deceive me? 00:15:28.080 --> 00:15:29.080 Dell wins? 00:15:32.080 --> 00:15:38.080 I mean, I know that a 4060 Ti is better than a 4060, 00:15:38.080 --> 00:15:42.080 but look at this margin of victory! 00:15:42.080 --> 00:15:45.080 And what a fantastic writing instrument! 00:15:45.080 --> 00:15:48.080 Get yours today at LTTstore.com. 00:15:48.080 --> 00:15:52.080 At 1080p, it's nearly a clean sweep. 00:15:52.080 --> 00:15:55.080 Red Dead Redemption 2, F124, 00:15:55.080 --> 00:15:59.080 Cyberpunk, Alan Wake, Dell demolishes the competition 00:15:59.080 --> 00:16:02.080 when it comes to pure gaming performance. 00:16:02.080 --> 00:16:05.080 It's not even really close. 00:16:05.080 --> 00:16:10.080 They do suffer one loss in Dota 2, where main gear's choice of a Ryzen processor, 00:16:10.080 --> 00:16:14.080 one that is cooled properly, not to mention quieter, 00:16:14.080 --> 00:16:18.080 helps them out in this CPU-bound game, and iBuyPower manages to pull ahead 00:16:18.080 --> 00:16:22.080 simply by, again, properly powering and cooling 00:16:22.080 --> 00:16:26.080 that same CPU Dell is using, but the rest of them, hey, 00:16:26.080 --> 00:16:29.080 a W is a W, and they're all Ws. 00:16:29.080 --> 00:16:34.080 As for our clear losers, Starforge lands at the bottom of the list, 00:16:34.080 --> 00:16:37.080 more often than not, with HP, main gear, and cyber power 00:16:37.080 --> 00:16:40.080 trading places for second and third worst. 00:16:40.080 --> 00:16:44.080 What's particularly interesting is the ray tracing testing we did. 00:16:44.080 --> 00:16:49.080 Yup, that's the RT performance that you can expect out of a 6650XT, alright? 00:16:49.080 --> 00:16:52.080 It was a modern card at one point, 00:16:52.080 --> 00:16:57.080 but AMD's newer cards are a lot better in this regard. 00:16:57.080 --> 00:17:02.080 At 1440p, it's more of the same. Dell loses to main gear and iBuyPower in Dota, 00:17:02.080 --> 00:17:08.080 and then crushes them, along with everyone else, in every other game. 00:17:08.080 --> 00:17:11.080 That 4060Ti is putting in the work. 00:17:11.080 --> 00:17:14.080 It's a real shame that our iBuyPower system didn't come with one, 00:17:14.080 --> 00:17:17.080 especially when it was supposed to. 00:17:17.080 --> 00:17:21.080 Whether the blame lies with iBuyPower, or with Amazon's crummy marketplace model, 00:17:21.080 --> 00:17:25.080 it has clearly hurt them here, as you'll see when we do our 00:17:25.080 --> 00:17:28.080 what could have been analysis later. 00:17:28.080 --> 00:17:32.080 Of course, raw performance doesn't tell the whole story. 00:17:32.080 --> 00:17:37.080 There's also the question of value, and here Dell's victory is still a victory, 00:17:37.080 --> 00:17:42.080 but not as clearly so. In Alan Wake 2, Cyberpower dethrones Dell, 00:17:42.080 --> 00:17:46.080 in Cyberpunk 2077, Dell wins by less than a dime, 00:17:46.080 --> 00:17:50.080 iBuyPower and main gear come out ahead in Dota 2, 00:17:50.080 --> 00:17:53.080 and then Dell wins by about a dollar in F124, 00:17:53.080 --> 00:17:56.080 and by about a quarter in Red Dead Redemption 2. 00:17:56.080 --> 00:17:59.080 Still though, what an upset! 00:17:59.080 --> 00:18:02.080 These results are really going to make the overall pick 00:18:02.080 --> 00:18:05.080 a tough call this year. 00:18:05.080 --> 00:18:08.080 Now, while Miss Katesen did specify 00:18:08.080 --> 00:18:13.080 that the system was for gaming, we thought we'd take a quick look at productivity performance. 00:18:13.080 --> 00:18:16.080 In our most GPU intensive test, Blender, 00:18:16.080 --> 00:18:22.080 StarForge is way down at the bottom, thanks to their lack of hip-RT capabilities. 00:18:22.080 --> 00:18:25.080 HIP is just AMD's version of CUDA rendering, 00:18:25.080 --> 00:18:30.080 and hip-RT is the latest version of it that utilizes ray tracing, but I digress. 00:18:30.080 --> 00:18:33.080 Dell is the clear winner here again, 00:18:33.080 --> 00:18:36.080 but that is their only win in productivity. 00:18:36.080 --> 00:18:40.080 In Cinebench 24, iBuyPower is on top in multi-core, 00:18:40.080 --> 00:18:45.080 and tied with Dell in single-core, while StarForge hangs out at the bottom. 00:18:45.080 --> 00:18:48.080 And in AV1 CPU encoding with Handbrake, 00:18:48.080 --> 00:18:53.080 Dell performs admirably, but loses to iBuyPower with StarForge on the bottom again. 00:18:53.080 --> 00:18:57.080 My detective skills are intuiting a pattern here, 00:18:57.080 --> 00:19:00.080 and X264 encoding tells us the same story 00:19:00.080 --> 00:19:04.080 as does PugetBench's Premiere and DaVinci Resolve benchmarks. 00:19:04.080 --> 00:19:08.080 Maingear manages to take second place in Photoshop, but other than that, their performance 00:19:08.080 --> 00:19:14.080 wasn't much to write home about. Yet, if we examine the cost of our systems, 00:19:14.080 --> 00:19:20.080 curiouser and curiouser, iBuyPower once again looks like a solid value proposition 00:19:20.080 --> 00:19:24.080 winning multiple rounds here. As for the worst value of the pack, 00:19:24.080 --> 00:19:28.080 HP, I'm sorry, but it's very clearly you. 00:19:28.080 --> 00:19:33.080 You are consistently at the bottom of every chart we're flipping through right now. 00:19:33.080 --> 00:19:36.080 Overall then, as surprising as it might be, 00:19:36.080 --> 00:19:40.080 Dell looks like the top pick whether you're looking at pure performance 00:19:40.080 --> 00:19:44.080 or performance per dollar. But value and performance 00:19:44.080 --> 00:19:48.080 aren't always number one in people's hearts. Some of you out there care about thermals 00:19:48.080 --> 00:19:52.080 or care about whether or not your power supply is reliable. 00:19:53.080 --> 00:19:58.080 To test our GPU thermal performance, we stressed our graphics cards with 4K F124 00:19:58.080 --> 00:20:01.080 and I'm happy to report that across the board, 00:20:01.080 --> 00:20:05.080 our system builders chose cases that do offer sufficient airflow, 00:20:05.080 --> 00:20:09.080 at least for the weak GPUs that we ended up with this round. 00:20:09.080 --> 00:20:12.080 Out of the whole bunch, StarForge performed the worst 00:20:12.080 --> 00:20:16.080 with still a totally adequate 75 degrees Celsius average 00:20:16.080 --> 00:20:20.080 and a max of 78. As for the rest of our builds, 00:20:20.080 --> 00:20:24.080 we're under 70 degrees on average with our coolest systems being the Dell 00:20:24.080 --> 00:20:27.080 and CyberPower at just 63 degrees average, 00:20:27.080 --> 00:20:30.080 though they were by far the loudest. 00:20:30.080 --> 00:20:34.080 As for CPU temperatures, things aren't quite as rosy 00:20:34.080 --> 00:20:38.080 for some of our competitors. Dell even managed to hit about 00:20:38.080 --> 00:20:42.080 70 degrees Celsius average during 4K gaming, 00:20:42.080 --> 00:20:46.080 which is a scenario that tends to be less demanding on the CPU. 00:20:46.080 --> 00:20:49.080 So what happens when we run Cinebench machines? 00:20:49.080 --> 00:20:53.080 Well, as it turns out, the boards really try to juice 00:20:53.080 --> 00:20:56.080 these processors, at least at first. 00:20:56.080 --> 00:21:00.080 iBuyPower and Dell were the only systems to reach over 200 00:21:00.080 --> 00:21:04.080 watts of power consumption during their Cinebench run, 00:21:04.080 --> 00:21:08.080 but they quickly realized that they can't handle that and throttle down to some pretty 00:21:08.080 --> 00:21:12.080 unimpressive speeds. Yes, 3GHz is higher 00:21:12.080 --> 00:21:16.080 than the advertised base frequency of this chip, but boy 00:21:16.080 --> 00:21:20.080 is it ever a lot slower than the maximum turbo, 00:21:20.080 --> 00:21:24.080 and Dell's average is even worse at just 2.88GHz. 00:21:24.080 --> 00:21:28.080 For those of you buying an iBuyPower system, there's clearly thermal 00:21:28.080 --> 00:21:32.080 headroom here based on our average temperatures, and you can adjust the power limits 00:21:32.080 --> 00:21:36.080 to get a little bit of extra performance out of your purchase, though 00:21:36.080 --> 00:21:40.080 that will take some tinkering in the BIOS, which you, the typical LTT viewer, 00:21:40.080 --> 00:21:44.080 might be inclined to do, but the average purchaser probably won't. 00:21:44.080 --> 00:21:48.080 As for our hottest system, it ended up being main gear 00:21:48.080 --> 00:21:52.080 at a sweltering 92 degrees celsius average, but 00:21:52.080 --> 00:21:56.080 to be fair to main gear, that CPU is pinned at over 00:21:56.080 --> 00:22:00.080 5.1GHz for the whole test, and like it or not, 00:22:00.080 --> 00:22:04.080 92 degrees is below AMD's TJ Max. 00:22:04.080 --> 00:22:08.080 So, I don't know about you guys, but I would rather have 00:22:08.080 --> 00:22:12.080 the hotter, more consistent running system over the one that's running cooler, 00:22:12.080 --> 00:22:16.080 but not living up to its potential. Either way, I think 00:22:16.080 --> 00:22:20.080 the win here has to go to Starforge. Whether it was their abundance of thermal 00:22:20.080 --> 00:22:24.080 paste, the contact frame on that Arctic cooler, or it just being 00:22:24.080 --> 00:22:28.080 an older CPU, it was only 700MHz from the max boost on average 00:22:28.080 --> 00:22:32.080 while sticking to around 51 degrees celsius. 00:22:32.080 --> 00:22:36.080 So this mediocre performing system will at least last you a really 00:22:36.080 --> 00:22:40.080 long time. Something else that might last you a long time 00:22:40.080 --> 00:22:44.080 is the 460W Platinum power supply from Dell. 00:22:44.080 --> 00:22:48.080 Don't get me wrong, it's only 460W and the proprietary connections means 00:22:48.080 --> 00:22:52.080 you can't build another computer with it, but when we ran through the light version 00:22:52.080 --> 00:22:56.080 of our PSU tester, we confirmed that it is highly efficient with 115V 00:22:56.080 --> 00:23:00.080 and 230V. We were so convinced that the results weren't 00:23:00.080 --> 00:23:04.080 correct that we went digging for the 80 plus test that confirms our findings. 00:23:04.080 --> 00:23:08.080 Incredible. iBuyPower used a high power 00:23:08.080 --> 00:23:12.080 J600GD F12S 600W. 00:23:12.080 --> 00:23:16.080 Very cool. This one tested poorly at 2% load efficiency, 00:23:16.080 --> 00:23:20.080 but this was otherwise a solid contender and mirrors our MSI and cooler master 00:23:20.080 --> 00:23:24.080 units for the rest of our load points. Speaking of which, Manegear and Starforge 00:23:24.080 --> 00:23:28.080 both used the same power supply. The MAG A650GL from 00:23:28.080 --> 00:23:32.080 MSI. These were the best power supplies out of the bunch, featuring 00:23:32.080 --> 00:23:36.080 good load regulation, efficiency rating and low voltage ripple or noise. 00:23:36.080 --> 00:23:40.080 Excellent selection. HP also chose well with their cooler master 00:23:40.080 --> 00:23:44.080 PSU. It matches the 80 plus gold rating and has mostly 00:23:44.080 --> 00:23:48.080 average voltage ripple and voltage regulation. A little high on the 3.3V 00:23:48.080 --> 00:23:52.080 but nothing to be concerned about. Which leaves us with cyber power 00:23:52.080 --> 00:23:56.080 and the Apevia ATX PR 600W. We didn't have a lot of faith 00:23:56.080 --> 00:24:00.080 in this one and while it didn't explode, we were right to doubt it. It's not 00:24:00.080 --> 00:24:04.080 efficient and doesn't stay above 80% efficiency at 100% 00:24:04.080 --> 00:24:08.080 load. We're not sure if this is just one bad unit or maybe 80 plus was sent 00:24:08.080 --> 00:24:12.080 a golden sample when they validated it for their rating. But it's not great. 00:24:12.080 --> 00:24:16.080 It's a clear tie here between Manegear and Starforge with HP as a runner up 00:24:16.080 --> 00:24:20.080 and I buy power with the honorable mention. If you buy from them, it looks 00:24:20.080 --> 00:24:24.080 like you'll get quality power supplies that should last you for years to come. And while 00:24:24.080 --> 00:24:28.080 the Apevia unit and the cyber power system isn't great, at least it uses standard 00:24:28.080 --> 00:24:32.080 parts. I'll take that over the low wattage, proprietary Dell concoction 00:24:32.080 --> 00:24:36.080 any day of the week, even if it is quite efficient indeed. 00:24:36.080 --> 00:24:40.080 Which leads us finally to a section we're calling 00:24:40.080 --> 00:24:44.080 alternate reality. No, not VR performance 00:24:44.080 --> 00:24:48.080 testing. VR hasn't been invented yet in my time. 00:24:48.080 --> 00:24:52.080 I'm talking about the timeline where we accept Manegear's offer, get 00:24:52.080 --> 00:24:56.080 the right PC for my buy power, don't have our order canceled with cyber power 00:24:56.080 --> 00:25:00.080 and spend our full budget with Starforge. 00:25:00.080 --> 00:25:04.080 Well, as you can see here, our productivity results don't change much. 00:25:04.080 --> 00:25:08.080 We mostly see the cyber power PC make some great strides, thanks to its 00:25:08.080 --> 00:25:12.080 stronger CPU, but these gaming results are something else 00:25:12.080 --> 00:25:16.080 entirely. It turns out that Starforge is the beneficiary of 00:25:16.080 --> 00:25:20.080 having a, yes, much much better GPU 00:25:20.080 --> 00:25:24.080 than the RTX 4060 Ti, allowing them to win in almost 00:25:24.080 --> 00:25:28.080 every test. With that said, when it comes to ray tracing 00:25:28.080 --> 00:25:32.080 which is becoming more and more important, it's astonishing 00:25:32.080 --> 00:25:36.080 just how much better those third gen RT cores are in NVIDIA's 00:25:36.080 --> 00:25:40.080 40 series GPUs. In Cyberpunk 2077, 1080p 00:25:40.080 --> 00:25:44.080 RT on, our Starforge 7700 00:25:44.080 --> 00:25:48.080 XT alternate reality system still manages to lose 00:25:48.080 --> 00:25:52.080 to our unchanged Dell XPS. And while the 4060 00:25:52.080 --> 00:25:56.080 Ti upgrade on some of our other systems certainly helps here, 00:25:56.080 --> 00:26:00.080 our Dell still manages to stay on top when it comes to 00:26:00.080 --> 00:26:04.080 pure value. Also, while we don't have the same 00:26:04.080 --> 00:26:08.080 120mm AIO that you would expect with the upgraded cyber power, 00:26:08.080 --> 00:26:12.080 take a look at these thermal results and please do 00:26:12.080 --> 00:26:16.080 yourselves a favor and avoid them. Unless you're buying a lower 00:26:16.080 --> 00:26:20.080 NCPU and trying to fit everything into a smaller case, these 00:26:20.080 --> 00:26:24.080 single fan AIOs are never worth buying thanks to their poor 00:26:24.080 --> 00:26:28.080 performance. You're better off either just strapping on a proficient 00:26:28.080 --> 00:26:32.080 air cooler or stepping up to a version with a 240mm radiator. 00:26:32.080 --> 00:26:36.080 With that segment out of the way then, it's time to declare the winner 00:26:36.080 --> 00:26:40.080 which will not be based on alternate reality, but only based 00:26:40.080 --> 00:26:44.080 on what we actually received. Well, Lylok, 00:26:44.080 --> 00:26:48.080 who was it? You may not like it, and I may not like 00:26:48.080 --> 00:26:52.080 it, but this is what peak performance looks 00:26:52.080 --> 00:26:56.080 like. No. Dell came first in gaming performance 00:26:56.080 --> 00:27:00.080 both in raw performance and in value, and they also 00:27:00.080 --> 00:27:04.080 won in our productivity suite and while performance isn't everything, 00:27:04.080 --> 00:27:08.080 it is a lot. The customer support could 00:27:08.080 --> 00:27:12.080 have been better as well as the ordering experience, but those are 00:27:12.080 --> 00:27:16.080 hopefully one-time events that you can suffer through one time. 00:27:16.080 --> 00:27:20.080 I would say the main knock against our little XPS 00:27:20.080 --> 00:27:24.080 that could here, and it is a big one, is that if you want something that not only 00:27:24.080 --> 00:27:28.080 looks like a gaming computer, but can be easily altered and upgraded 00:27:28.080 --> 00:27:32.080 and tinkered with like a gaming computer, good luck doing anything 00:27:32.080 --> 00:27:36.080 with this proprietary board, weird CPU cooler mounting, and 00:27:36.080 --> 00:27:40.080 460W power supply that, frankly, we were amazed 00:27:40.080 --> 00:27:44.080 didn't blow up when Constable Lucas was testing it. With all 00:27:44.080 --> 00:27:48.080 of that said, as a set and forget gaming box, darn it, 00:27:48.080 --> 00:27:52.080 I just can't argue with the evidence. For a runner-up 00:27:52.080 --> 00:27:56.080 and system that I would be more likely to pick, it's a tough call. 00:27:56.080 --> 00:28:00.080 If we'd received the right system for my bi-power 00:28:00.080 --> 00:28:04.080 directly, they'd probably take the silver medal, but that loose motherboard 00:28:04.080 --> 00:28:08.080 and the dodgy purchase experience ... 00:28:08.080 --> 00:28:12.080 As for Starforge, if we'd gotten the 7700XT, there's a solid chance 00:28:12.080 --> 00:28:16.080 they would have won the whole thing, but based on what we actually 00:28:16.080 --> 00:28:20.080 got this year, I think second place has to go to cyber power. 00:28:20.080 --> 00:28:24.080 Purchasing from them sucked. What are you calling for? 00:28:24.080 --> 00:28:28.080 What are you calling for? I want to buy the f***ing computer! 00:28:28.080 --> 00:28:32.080 But customer support wasn't bad, even if they just directed us to a video 00:28:32.080 --> 00:28:36.080 and at $1100, it's really hard to beat this machine. 00:28:36.080 --> 00:28:40.080 I guess you're right, but I have to say that this year just didn't 00:28:40.080 --> 00:28:44.080 excite me as much as it normally does. These new fangled computer gadgets 00:28:44.080 --> 00:28:48.080 seem to cost more and more each year and these 60 class cards 00:28:48.080 --> 00:28:52.080 seem to go less and less far each year. 00:28:52.080 --> 00:28:56.080 What do you say next year, or two years? We up the ante. 00:28:56.080 --> 00:29:00.080 Let's butter our bacon and buy some $2,000 gaming doohickeys. 00:29:00.080 --> 00:29:04.080 Excellent. We'll need a sponsor though. Thank you all so much for watching. 00:29:04.080 --> 00:29:08.080 If you want more Secret Shopper, check out the last time we did, I can't believe Del won this year. 00:29:08.080 --> 00:29:12.080 It's first time for everything. But that? It's first time for everything. 00:29:12.080 --> 00:29:16.080 I don't know how they did it. He walked into my office for script review, 00:29:16.080 --> 00:29:20.080 okay? And he goes, you're not going to believe who won. How many guesses did it take me? 00:29:20.080 --> 00:29:24.080 Four or five? At least four. 00:29:24.080 --> 00:29:26.080 He did not want to believe it was Del.