MSI Fans should protest the BGA CPUs. You're paying for 980M with a 4720HQ? Or even 980M SLI?
This issue I see is MSI prices have been skyrocketing. They were when I bought the Asus G73JH (First gen) the budget machine, and then just a smidge over P150EM later. But now the 980M GT72 is $2200? The 980M SLI is well over $3000.
Just throwing it out there, I think paying $2200 or more for a system with BGA 4720HQ is a travesty, it's insane.
What's the difference? If you're into encoding video example, 4720HQ by a NBR member's test was 13.8 FPS. My 4790K averaging 20 FPS. That's a difference of 26 minutes for 72000 frames, 4790K completing in 1 hour vs 1:28.
Imagine if you're encoding a quality rip that is about 7 FPS on 4720HQ for 3 hour encode, but 12 FPS on 4790K, that's a ton of time you can save. 3 hours vs 1.75 hours.
-
Yeah good luck with that one. Say the entire notebookreview MSI members say they no longer will buy a MSI machine again.
I bet you MSI make millions of dollars on the rest of buyers that dont care, on future sales of new complete systems where the CPU died on their previous notebook or they wanted the newest and fastest CPU from Intel.Spartan@HIDevolution and Talon like this. -
-
Well, that is mostly a decision by intel. You are comparing a mobile CPU with half the TDP to a desktop, practically top of the line quad core from intel. I am not concerned with 4720HQ. It performs excellently in gaming and I have zero issues so far. What I am concerned is with the TDP limit for a more powerful CPU, like the 4980HQ. That CPU, if it was 57w, it would perform much faster than its current incarnation.
Anyways, I'd say the mobile CPUs are no more socketed, so laptops like the GT80 titan should probably go the desktop CPU route.Charles P. Jefferies and Starlight5 like this. -
Yeah. I spoke to Eurocom yesterday. I've also been up for about 24 hours. Anyway.
Eurocom. Right. I told them about the terrible HQ CPU trend and asked if they had any clout with Intel to return to socketed. He said that the next 2 gens appear to be locked in for HQ only, but that he definitely would be interested in telling him about the customers that want sockets back.
So... contact MSI. Contact Eurocom. Contact Sager. Contact ASUS. Contact everyone you feel comfortable contacting, and express the desire for socketed chips/models to return (make sure you want CHOICE, not all sockets) and we might get somewhere.TomJGX, Starlight5, TBoneSan and 1 other person like this. -
thegreatsquare Notebook Deity
Luckily I got my GT72 for $2k on sale. I'm exclusively gaming, so I'm not sweating performance.
I do think BGA should have a better warranty, at least 3 years ...even if just for the mobo/CPU itself.TomJGX, Starlight5 and SeagateBoy like this. -
nightingale Notebook Evangelist
While i agree with you that bga is absolute turd, intel no longer produces socketed cpus for mobile as far as i know.
On top of that, the fact that ever since sandybridge the processor performance have been sidegrades or miniscule improvements at all if any, it really calls to question do we need new cpu's. the 4720 will most definitely outlive the 980m for the foreseeable "gaming future".
MSI has done us a great favour of giving us mxm upgradeable graphics interface, in comparison to asus or the 2015 alienwares and razer etc who are all using soldered cpu AND gpu which will definitely mean the laptop will become outdated soon enough when the 900m series no longer cuts it for top gaming.Starlight5 likes this. -
120Hz CPU-heavy gaming? No.
Render machines? No.
Heavy video editing work? No.
Livestreaming machine? No.
There's a lot more than gaming, and enthusiasts get machines for more than gaming, which is why those chips are such turds. They can't clock up and hold it under heavy load. You always have to fit into the 47W TDP. Even for the 4980HQ which can clock to 4.2GHz on 4 cores.
As for the non-MXM, soldered stuff, it's because if you don't solder, you don't get thin. Thin = solder to board. They went for thinner machines, so they had to solder. Most people don't care because thin > everything else they need, apparently.Kent T, TBoneSan and Starlight5 like this. -
Having a choice is nice. Though what I particularly would like is TDP control.
Even at 4Ghz, you are seeing merely 10%-15% performance difference. If a 4Ghz CPU barely reaches 120fps, guess where I will be? around the 100s, and that is on CPU bound games.
As for the rest, sure, time can be exponential when rendering, but then again, talking about laptops, it won't be nearly as massive as vs a desktop system. You will just wait a bit more time. There are very few models that offer more performance than a 4720HQ, while a ton more that offer less, or much much less. Rendering, Heavy video editing, livestreaming, all those things can easily be done here. Are there better options out there? Sure, but laptopwise we are limited, and this is, in the end, a gaming machine.
Personally, I would have preferred a 4980hQ with higher TDP. It hardly makes sense to put such turbos on these haswell CPUs, including socketed ones, with such a low TDP. 4980HQ can consume upwards of 80w for 4+ghzso naturally it can't hold turbo that high.
I can't speak of any other 4720HQ machines though. I know some that perform very low, as seen with passmark scores being all over the place... Some score barely 7k, mine is shy from 10k. Huge difference for the same CPU.Charles P. Jefferies likes this. -
The 4720HQ is a 2.6GHz base, 3.4GHz 4-core turbo chip. It can be OC'd to 3.6GHz under 4 cores. If you are using 2 cores it will turbo to 3.5GHz and 1 core to 3.6GHz naturally, without overclocking. OCing can bring that to 3.7GHz and 3.8GHz respectively.
A CPU hitting 100% is not indicative of how much TDP it is drawing. For example:
Intel XTU bench pulls ~60-66W. CPU is set to 3.9GHz, stock voltage, 100A limit on my 4800MQ.
Handbrake rendering Shadowplay videos of Battlefield 4 (1080p, 60fps, "high" quality for Shadowplay settings) using "veryfast", "main", "constant FPS @ 60" consumes 68-75W while rendering.
LinX Linpack set to 4 cores rendering will pull well over 90W... while throttling due to current (if on adaptive voltage). It would theoretically pull even more if I set static voltage at ~1.2v.
Playing Dark Souls 2 + livestreaming with some heavy compression on: ~60W+
Pre-patch BF4 at 125fps limit? 60W+
Then on the contrary
Mass Effect 3 at max graphics, 125fps limit, 4x MSAA + 2x SSAA forced via nVidia Inspector. Power drawn? 42W.
GRID 2 at max graphics unlocked FPS? ~40W
Post patch (yesterday's patch) BF4 (where they actually optimized the bloody thing) ~45W, spiking to 50W with tons of action on-screen.
And most all of the gaming-related incidents I listed used similar amounts of CPU power too. ~70-80% on average. I also watch a lot of livestreams while I'm gaming, so that's extra load splitting & power draw =D.
But I wasn't wrong. For a user who needs to do CPU-heavy things, like streaming/lots of rendering/etc, the HQ chips are pretty bad, because they don't go beyond 47W if under load for long periods of time.
Now, as I said, for the pure GAMER, it's much less of an issue. Especially for the 60fps gamer; it won't matter one bit I'm sure.
But there's a lot of things that are different for the gamer and the multitasker/enthusiast. I happen to be the latter. Heck, even overclocker-benchers would get by on the 2.5 minute time limit for excess power draw as most CPU-focused benches are within 1 minute in length.
Anyway, the 4980HQ isn't much different at stock than your 4720HQ. It only hits 3.6GHz at 4 cores. Unlike the 47xxHQ and 4xxxMQ chips, the 49xxHQ line's 2-core turbo is 200MHz above its 4-core and its 1 core is 200MHz above its 2-core.
So a normal chip would go: 3.6GHz 3 & 4 cores, 3.7GHz 2 cores, 3.8GHz 1 core.
49xxHQ would go: 3.6GHz 3 & 4 cores, 3.8GHz 2 cores, 4GHz 1 core. It's why I keep harassing reviewers to list what the actual 4-core boost is and not just the 1-core max boost listed by Intel.
It is true however that overclocking it has the +600MHz limit that the 4900MQ and 4910MQ chips enjoy, which is probably what you meant when you said TDP control should be a thing.
As for the 4GHz thing, 10% extra performance in certain things is a huge deal. I can say that the difference between 3.3GHz and 3.5GHz while streaming is using "medium" or "fast" in certain games, or using medium/fast and having more breathing room for the game itself. There's a similar difference between 3.5GHz and 3.9GHz like I currently overclock to, and there'll be even more of it between 3.9GHz and 4.3GHz when I acquire a 4910MQ chip (whenever that is). CPU power is more important to me than it is to many others, but again, the fact that these chips with those limitations are what's being forced on us is the real problem. Anytime someone says they wanna game, I say okay cool, enjoy. When someone says they wanna render/3D render/etc I absolutely must point them to a MQ chip that can be adjusted properly in XTU. Especially if it's a machine in place of a desktop they're getting.
Plus, most of these things are Desktop Replacement laptops. My D900F was and my P370SM3 is, and the higher end stuff like your GT72 also is. Those things are designed to and should work 100% how you demand them =D.
In conclusion to the book I wrote, all I'm saying is, if we're forced to use x kinds of chips, then x kinds of chips should be functionally equivalent to the high-end stuff we had before. Even in the case where Ivy Bridge is superior to haswell in power draw and heat output, haswell still functioned properly for the most part, right?
Anyway, all that aside, I WILL point out that undervolting the CPU does help. Less volts means less drawn power which means higher clocks under same load. I did find a guy who managed to undervolt so well that he could set his CPU to 3.6GHz and never throttle in XTU benching etc. He was a lucky one.
Okay, I've said enough. I'ma stop talking now. #Perfectionist -
I agree and understand what you mean, and I also share the dismay with the TDP limit. Though, as you pointed out, my 4720HQ is not nearly as hurt as the higher end versions which just end up with wasted potential.
D2 Ultima likes this. -
thegreatsquare Notebook Deity
-
nightingale Notebook Evangelist
More or less, this is about expectations. You cannot expect a gaming orientated laptop to perform outstandingly for workstation projects, different animals for different situations. -
-
nightingale Notebook Evangelist
D2 Ultima likes this. -
Don't question my specifics for slap locations. Discussions in high school were weird.nightingale likes this. -
nightingale Notebook Evangelist
-
The problem is how they market things, really. They will shield themselves for now with the excuse that their CPUs are really 2.6ghz quad core, and the turbo is not guaranteed or something like that.
In reality they just need to sell more options, with full customization for TDP, clocks, all. Like their K versions on the desktop side. As things are, their whole HQ line is basically meaningless (sans for the select few that use iris pro) because they, over time, regulate to the same TDP and end up with the same 4720HQ performance.
There was a slight shimmer of light recently though, the unlocked GT80 BIOS seem to have options for TDP, core clocks, ratios all etc on bios. With luck, maybe we can tweak these HQ nonsense processors to shape so they can become powerful enough. While gaming will be alright for these, it is a shame to have lower potential performance.... -
There are applications where a soldered CPU and GPU make sense, such as in the Sager NP8651 and MSI GS60 Ghost. When you choose one of those laptops, you are choosing to sacrifice upgradeability and a little CPU performance for increased portability. For someone like me, who has to carry their laptop around a lot, that compromise is worth it. In those cases, there is a logical reason why someone would want a soldered CPU over a socketed one.
But putting soldered components in desktop replacements, like Alienwares or the MSI GT line, where you don't gain anything in return? Yeah, that is stupid. At least Clevo/Sager has continued to offer socketed CPUs in its top end laptops. -
The same thing seems to happen on MSI's MQ chips too, but this is presumably on purpose because their MQ-using models have a 180W power brick and sold with 100W-125W TDP chips (680M, 780M, 880M) prior to the 980Ms. Even Svet's unlocking doesn't let them draw extra power.
I wanted an AW17 R2 user to check with a 4980HQ to be 10000% sure, but those guys are so hard to come across. -
I never saw the benefit of MXM personally (I'm not saying it's not greatly beneficial for other people) as it wasn't worth the cost of upgrading my old P170SM to a 980M. The P651SG had far superior cooling in a much more compact chassis, with far superior IPS screen options available and better speakers. It was a no-brainer to get the entirely new system instead of trying to import a 980M from foreign lands (there is no MXM 980M available from any store in the UK that I can find through Google) and deal with the outdated P170SM chassis. Even if I could find an available listing somewhere, going by historical pricing patterns the MXM 980M's would probably amount to half the cost of an entirely new system anyway! Such is the horrific technology price gouging we have to endure here in the UK.
Same story with socketed CPUs - the i7-4910MQ is currently £450 right now! That is insanely expensive. Around third of the cost of an entirely new laptop with i7 & 980M combo. The i7-4810MQ is £322. Better, but still vastly overpriced compared with the desktop CPU counterparts. i7-4710MQ is seemingly unavailable currently.
I like the idea of having socketed components but the ridiculous price gouging in the UK and general lack of availability in this country (especially in the case of MXM GPUs) make socketed laptops almost entirely redundant. Sad, but true. -
So, while I love building PC's and love laptops, BGA's are frustrating but aren't the end of the world to me in an awesome laptop, and aren't MSI's issue.... Intel is doing this, so fuss to them! The prices of mobile CPU's and MXM upgrades ruined this area long ago!!Cakefish likes this. -
How does the P650SG have better cooling than a P170SM? You would need to get a P170SM and put your P650SG next to it and clock the CPUs the same and the GPUs the same using the same thermal paste on each model and then test. The P170SM and P170SM-A have MUCH better CPU cooling for sure. GPU cooling might be a stalemate, due to the dual-fan setup of the P6xxSx models for the GPU cooling systems.
As for the screen, there are quite good TN panels for the P170SM that you may not have had, but it's not to say that they didn't exist. The 90% gamut AUO panel is one of the best ones actually, and should have better colours and has better response time than the IPS you have.
I give way to the speakers, Clevo speakers have sucked for ages upon ages. -
Heh, I think the IPS displays due to response suck for gaming. 25ms for a gaming panel? If you suggested that as a gaming panel at any other forum, they would likely laugh and insult it or just think you're a troll.
-
-
HTWingNut likes this. -
-
If we add on the price of a screen upgrade then it would have made even less sense to upgrade the old machine vs buying an entirely new one. I'm not saying my situation applies to everyone around the world but for the UK it just doesn't make sense to buy all these individual socketed components, due to how expensive they all are - and their extreme scarcity.
-
Also, MX-4 sucks. =D. I've not seen users attempt PK3 on my machine, but believe me when I say MX-4 should be ignored for laptop usage. MX-2 as well. -
OK, alright... I'm willing to believe that Maxwell is just that good. I stand by the rest of my points though!
Yeah, PK-3 did offer better results than MX-4 from my experience. -
-
-
MSI Fans Should Protest BGA
Discussion in 'Gaming (Software and Graphics Cards)' started by Zymphad, Mar 3, 2015.