I present to you Firebat's Blade 15! (144hz/i7-8750H)

So, in light of your question, I just ran 6 back to back to back benchmark runs of Cinebench R15, with HWInfo64 running in the foreground so I could watch the output in real-time.

I do use multiple profiles in Throttlestop to maintain temps while gaming, but one of my profiles (default, actually) is bone-stock (no undervolt, no turbo limits, nothing...).

Across 6 back to back runs, I saw all 6 cores boosting up to 3.9 GHz, and CPU Core #2 did hit 99C (rest were 91C or below). So my one core did thermal throttle.

When I check power, my CPU hit a peak package draw of 57.281W.

No repasting or anything done to mine, and in this case, it was running completely stock settings, Windows Balanced Power mode, and Synapse 3 set to Gaming Performance mode.

So in my case, I can hit boost clocks and stay there the entire time Cinebench R15 is running, and it seems to be pulling full power.



Patrick

 

I am getting power limit throttled no matter what combination of settings I run/tried with balanced/gaming mode... Undervolting... repaste/new thermal pads etc. It is something Razer engineers are working on and aware of. Hopefully we will see a fix sooner than later. I will keep everyone updated on any important information in this thread. I would not jump to RMA yet... as this fix would involve a Bios or EC update and should not be hardware related. I will be testing out a third Blade this weekend to see what kind of results I get. I'm almost positive this power cap is across the board, though.

I have the same experience... however only when JUST stressing the CPU. Once I stress CPU and GPU I will get either power limit or thermal throttling even if the CPU temps are in the 70s-80s. I can assume you that you will see the same issues in gaming situations, depending on the game. When my CPU is in the 80s during multiple cinebench runs.. it doesn't throttle at all. Same temps... but with GPU load.. throttling.

 

Ahh, now I see what you're experiencing. I was just playing Doom for a bit (using my "Gaming" profile in Throttlestop). And I noticed that my CPU power draw never exceeded 36W. I forgot to check what my boost ratio was, but given the lower power draw, I'm pretty confident that it wasn't boosting up to 3.6 GHz (my peak turbo ratio under TS for my "Gaming" profile).


Patrick

 

I am actually getting a replacement Blade in the mail later today... mostly because of 3 stripped threads so my back panel wasn't being held on very well. I will let everyone know how performance is from this latest model. Also I know the Razer engineer team is working on the throttling issue so it's definitely not going unnoticed.

 

So if they fix this issue, ideally we should be getting 3.9 GHz with all 6 cores running indefinitely (at 45 W TDP minimum) without any throttling right?

The first blade 15 I got on late May had bad thermals (reaching 100 C) and severe power throttling on numerous tests despite undervolting.

My current blade (manufactured July) on its first fire strike test got intermittent thermal throttling throughout and a power throttling spike just as the physics test portion ends, but out-of-the-box score is 14,158 with max CPU package temp of 95 C, max GPU temp of 73 C (already better than the undervolted first blade). It is summer and I don't have air conditioning so I guess it's not that bad? This is on synapse gaming mode at 5k rpm fans.


Setting CPU core & cache to -133mV, speed shift to zero, GPU overclock (core +200, memory +600 MHz), I got improvements to score (14,718) and temperatures (max CPU package at 90 C, GPU at 70 C) and zero thermal/power throttling.


Cinebench runs (scores range from 1079 to 1182) are a bit short so I also did CPU stress tests on XTU.
Still at -133mV, and default max turbo of 3.9 GHz for 6 active cores:

cinebench CPU test run:

synapse gaming mode:

3.88 GHz @ 66 W
eventual power throttle (just before the run ends) to:
3.25 GHz @ 45 W
max package temp: 84 C​

synapse balanced mode:

immediate power throttle to:
3.25 GHz @ 45 W
eventual power throttle to:
2.75 GHz @ 35 W
max package temp: 70 C​

XTU CPU stress test:

gaming:

3.85 GHz @ 55 W
eventual intermittent power throttle (after ~1 min) to:
3.6 GHz @ 45 W
max package temp: 81 C​

balanced:

immediate intermittent power throttle to:
3.6 GHz @ 45 W
eventual power throttle to:
3.1 GHz @ 35 W
max package temp: 73 C​

testing with max turbo lowered to 3.6 GHz for 6 active cores:

cinebench CPU test run:

gaming:

3.58 GHz @ 55 W
eventual power throttle (sometimes the run ends fast enough before throttling) to:
3.25 GHz @ 45 W
max package temp: 78 C​

balanced:

immediate power throttle to:
3.24 GHz @ 45 W
eventual power throttle to:
2.8 GHz @ 35 W
max package temp: 70 C​

XTU CPU stress test:

gaming:

3.58 GHz @ 45 W
no throttling
max package temp: 73 C​

balanced:

immediate intermittent power throttle to:
3.4 GHz @ 40 W
eventual intermittent power throttle to:
3.0 GHz @ 35 W
max package temp: 70 C​

Also note these are all with stock thermal paste (won't repaste if I don't have to) and fans set to 5k rpm. Normally I just set speed shift to 128, and lower the fans to 4.8k rpm with synapse gaming; automatic when balanced.

 

Anything under 4,1 is actually a throttling on 8750H, they seem lock it to 3,9 for better thermal. Gaming mode is set to throttle to 45Watt and balanced mode to 35 Watt after around 10 second, for full cpu performance I have to use gaming mode. Though it’s not a deal breaker to me, and you have good temperature btw imo.

 

These aren't bad scores at all.. but to answer your question. Yes without power limit throttling you should be able to maintain 3.9 on 6 cores at all times. Hopefully it gets fixed soon!

I wouldn't call 3.9 throttling... that is normal for running on all 6 cores. The 4.1 really is if you are only running 4 cores.

 

Yes, it’s true i read also on all cores it should be 3,9 but intel stated it up to 4,1 turbo boost so I was hoping to see this frequency, but never saw that, so to me it feels throttling or kind of bios limiting and other manufacturers are limited that high frequency apparently to maintain temperature on this chip.

 

Actually, I think it will only hit 4.1 GHz if you're running on two or less cores. Running 4 cores the boost max drops to 4.0 GHz.

I suppose an easy way to test would be to run Cinebench 15 in a single core test, and see if it ramps up to 4.1 GHz on it's own. That would at least prove that the system is capable of it. I wonder if there are any benchmark tools that allow you to restrict the number of cores in use (beyond just the single core limit)?



Patrick

 

Yes you are correct lol sorry its 6 cores at 3.9, 4 cores at 4, and 1-2 cores at 4.1 I believe.

 

So it’s marketing “throttling” who use only 4 cores in 6 cores cpu, I think no one, intel should marketing it 3,9 because casual user only use as it is and take a look at value that they advertised, or stated boost up to 4,1 on single core usage.

 

Actually most games don't even utilize 6 cores.. they are only using 4. I tried disabling 2 of the cores to see if I could run 4ghz at 4.. but all 6 were still being "used".

 

I see. Btw how Far Cry 5 runs now? It’s maybe because summer temperature over 30 here CPU now hits high 90ish, I hate looking at HWINFO now Though game runs fine, but I feel like i barbecuing my Blade lol

 

Farcry 5, FFXIV, Battlefield 1, and Assassins Creed Origins all get to the 90s and sometimes hit 100c for me. I'm not even gonna bother repasting again.. stock vs Thermal Grizz was not worth it. The Graphite based paste Razer uses is actually pretty good. Unless you are gonna LM it (which we still haven't even seen great results from) I wouldn't even touch it.

 

Do you still have Blade 14 7700HQ? Since I don’t have it anymore, when I’m not wrong I didn’t hit 90 on that Blade, I played also Battlefield 1 and AC Origins on that. Yes you’re right I don’t have a plan repasting my Blade atm, don’t see great result yet, I think they do right job with cooling system now, stress test is fine on max 70ish only cpu temperature on games and maybe SSD on default are high wich can easily be done with thermal pad, though I’m too lazy doing that

 

I don't exactly have it anymore.. I sold it to my best friend. He's not into all the benchmarking etc so it's not even like I could ask him to do it, haha.

 

So now that I've had this system for over two weeks, I think I'm mostly done with the benchmarking / testing / tweaking side of it.

I was running MSI Afterburner to see if I can squeeze some extra score in FS, but I realized that I wasn't really getting any better scores. With the stock (Razer) 389.04 drivers, I'm getting scores of 13,917 running stock config on my Blade 15 (no undervolt, no turbo boost limits, etc). I had noticed earlier on that I had gotten some scores into the mid 14,xxx (peak of 14,445) mark, but I've never been able to reproduce that score. MSI Afterburner, even running +160 Core (max I could get to run stable) and +699 Memory, I only got up to 13,940 on FS. So that leads me to believe there is some significant power throttling going on, if a significant overclock on Afterburner isn't getting me anything in FS or Heaven (Heaven, for some reason, still seemed to report stock Core / Memory speeds).

While I have been tempted to go back to XTU to see if I can improve the numbers at all, in reality, there is only so much benchmarking one can do....at some point I want to just use the machine for work, and to play some games on it. And I do like how I can create auto-loading profiles within Throttlestop for conditions (ie, I can create a profile that it will switch to if my CPU peaks above 90C, for example, and I have a specific profile for battery power mode that disables Turbo boost, and gives me 8+ hours of battery life).

As to repasting, I've thought about this myself, but temps aren't bad, and I can limit them to below 90C with a Throttlestop profile setup for gaming (-0.125mV undervolt and Turbo Boost max at 3.6 GHz vs 3.9 GHz). Doom, Crysis 3, and Battlefield 1 all keep the temps below 90C, and no thermal throttling of any cores. If I had more confidence in a liquid metal repaste I might try that, but given the risks, it's just not worth it. Any other repaste is likely to just result in basically the same scores, which tends to make me think Razer did a reasonably good job on the initial paste on both CPU and GPU.

I did throw a couple of Silverstone thermal pads onto the SSD (2TB M.2 Samsung 970 EVO), as my temps when playing games rose up to 89C occasionally on the SSD itself. With the thermal pads in place, I have not seen a temp on my SSD above 53C.

I have to say, about the only complaint I have about this machine is that the keyboard can sometimes feel "slow", like I missed a key or something. All of the keys work perfectly fine, but I occasionally wonder if I'm almost typing faster than the keyboard can keep up. I do like the feel of the keyboard just fine, and it's far, far better than the Apple MBP keyboards of the past year or two. And it's still not anywhere near the keyboard quality that you'd get on a Thinkpad X1 Carbon, for example (still my favourite laptop keyboard).

Otherwise, I'm pretty much thrilled with this machine. I just need Razer to get some spare 230W power supplies in stock so that I can leave one in my home office, and one in my actual office, and then I don't have to pull my big power brick out of my laptop bag unless I'm remote.



Patrick

 

Seems like everything is in check with your system and I'm glad to hear that! You really don't need to repaste based on your details. I would love to see some pics of 30 min gaming sessions to show 0 throttling with your system. I am trying to collect as much data on this as possible.

 

For those want some repast data, I did a repast with kryonaut. On games like hearthstone before repaste the temperature stay around mid 70s after repaste it stay around mid 60s. On the Witcher 3 before repaste it was around mid 90s highest 100, after repaste, it stay in mid 80s highest 93. All undervolt -140 on ts. So for me personally the repaste did pay off. I’m not planning to paste lm tho. Hope this helps

 

I am looking for anyone that is experiencing power limit throttling on their Blade 15.. I may make a new thread about it but for now am posting this here. I need at the very least screens showing whats happening so I can collect data to forward to to the team. Thank you all in advance!