Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
iPhone 11 iPhone 11/Pro/Max 2.66 GHz, 4 GB RAM - Geekbench 4 >14000, Geekbench 5 >3500
- Thread starter EugW
- Start date
- Sort by reaction score
Unlikely.. I’d like it if they did obviously, but this was all a wild fire of speculation. I pre ordered a damn Pro Max 11 256Gigger based on thinking it actually had 6GB of ram “sounded to good to be true” . And I didn’t read further at all about this. I just saw some incorrect specs posted up on retailers websites and took it for the real thing.
I’m going from the X to 11 Pro. I think it will be a nice upgrade.
IMHO, no, for 99% of XS Max owners.
I already posted this but I’ll post it again. 4 GB RAM in 512 GB Pro Max:
Hi, I have an XS Max. What do you think? Could I run the new game which was introduced on the Keynote - Pascal’s Wager? Is it possible that I could run this game on the highest graphics setting on the A12 Bionic?
No A13 is about 20% faster. It was only the first leaked A13 scores that didn't look so good. Subsequent ones were much better though. Geekbench 4 is about 13800 with A13, and Geekbench 5 is about 3500.
My understanding was that the carriers don't even get the RAM specs. Apple itself doesn't even list the RAM specs anywhere on their own website.
I know that in the past carriers in various regions have had specs listed that simply turned out to be wrong. It would seem that some overzealous data entry people put in specs into the database that are just best-guess numbers.
What would account for the A13 appearing to be slower in early tests? And where are these phones coming from, a week ahead of official release?
[doublepost=1568471204][/doublepost]
Now *that* is a surprise..
Early revisions of CPU silicon don't perform as highly as final revisions. Software not optimized yet can account for lower performance as well.
Geekbench shows a lot of variation. If you look at the listed score charts for the listed iOS devices, you'll see that the scores listed in those charts are actually much lower than what individual devices can achieve.
Why? Because there is a lot of variation in the scores, so for the purpose of the chart they take a weighted average or something like that. However, if you really want to see what the devices can do, you're better off looking at the individual scores, and throwing out the bottom 90% of the scores.
BTW, if the phone is warm and it's really warm in the room, the phone may score lower due to throttling.
So, maybe different versions of the iOS 13 beta? On the silicon, how often does that get revised on a device like the iPhone (in general)? I've sometimes wondered whether the early phones were not as good performers as phones manufactured three months later, but have no data at all to support that.
[doublepost=1568471882][/doublepost]
I guess that's not surprising. I used to be a serious ham and there was a fair bit of variation unit to unit.
And the temperature makes a lot of sense. Another scenario would be where the phone was just charged, in which case everything inside could be somewhat warmer.
This guy knows about the same as us.
Apparently he didn’t follow EugW thread. Hell, I didn’t either. I was reading the thread, and stayed off macrumors for a couple days. Then I pre-ordered a promax, then I read the latest post’s in this thread and realized how misinformed everyone really was.
For my 12" MacBook, another fanless device, I would consistently get higher repeated scores in Cinebench if the MacBook was sitting on a granite countertop, as compared to the MacBook on a wood table. It's presumably because granite acts as a heatsink, whereas wood acts as an insulator. Both were in the same room at the same room temperature.
https://forums.macrumors.com/thread...uns-of-cinebench.2073415/page-2#post-25271927
For this test, we MacRumors members ran Cinebench repeatedly, and recorded the score each time. Run number on the x-axis, and score on the y-axis. The m3 scores are my own. (The i5 and i7 scores are from other people.)
I think it’s worth noting, anyone can say anything about what their ‘score’ was, why don’t you post the results for everyone else to see in comparison, (assuming you don’t reply), then we can safely assume that you might be misleading others based off your claim.
That said, the A13 is actually showing scores higher over the A12.
He's correct, at least regarding the first A13 scores. A12 multi-core scores were higher than the first A13 multi-core scores. These A12 scores are well-documented on the Geekbench website.
However, subsequent A13 scores are significantly higher, also well-documented on the Geekbench website.
Last edited:
Mobile devices love to throttle. I imagine if you put liquid metal on the 12” MacBook it would score drastically higher, as the temps would probably instantly drop 10-20C.
Especially running cinebench. Your multithreaded score will consistently drop after every loop until the device reaches maximum temperature. Or average throttle.
[doublepost=1568475613][/doublepost]
Damn that thing is fast. Woot woot! Can’t wait to get mine.
Especially running cinebench. Your multithreaded score will consistently drop after every loop until the device reaches maximum temperature. Or average throttle.
[doublepost=1568475613][/doublepost]
Damn that thing is fast. Woot woot! Can’t wait to get mine.
That’s great! We really need to clear up this unknown about the RAM in the Pro. Concerning to see 6 Gb advertised in places but Geekbench showing 4 Gb.
So I am crazy and skimmed through all 114 Geekbench entries (98 GB5, 16 GB4) now in the database.
1. All are 4 GB. However, iPhone 11 is 3844 MB, iPhone 11 Pro is 3759 MB, and iPhone 11 Pro Max is 3740 MB. Is this because of the RAM allocated to graphics? iPhone 11 would need the least RAM, iPhone 11 Pro would be in the middle, and iPhone 11 Pro Max would need the most.
2. Most of the entries are 2.66 GHz. However, two entries are for 1.86 GHz. I did some quick calculations and I'm wondering if the actual clock speed is not 2.65 GHz but in fact is 2.6666... GHz. If that 1.86 is actually 1.8666..., then that is EXACTLY 70% of full clockspeed. So I wonder if it's low power mode, which in this case cuts 30% off the clock speed.
3. At 1.86 GHz, the score is 783 / 1560 for Geekbench 5. That 30% number I quoted above doesn't make sense here in terms of performance, since that's only about half of expected performance. What could account for this? Is there a different mix of performance vs. efficiency cores at play here? Note that for 1.86 GHz, it still lists 1 processor with 6 cores like it does at 2.66 GHz, but at 1.86 Hz, the multi-core vs. single-core multiplier is 2:1. At 2.66 GHz, the multiplier is more like 2.6 or so.
4. There is indeed much variation in scores. Some of the worst scores include both the 11 and the Pros, and some of the best scores include both the 11 and the Pros. So, performance is not linked to model. The comparatively crappy scores we saw early on for iPhone 11 were most likely just because of variation in this benchmark, +/- throttling.
5. Most of the variation in scores has to do with multi-core. Single-core performance is mostly consistent, but the multi-core ranges much more from about 2200 to about 3550. I wonder if throttling is involved here, since presumably throttling would affect multi-core performance more than single-core, because of the higher heat generated from multi-core.
1. All are 4 GB. However, iPhone 11 is 3844 MB, iPhone 11 Pro is 3759 MB, and iPhone 11 Pro Max is 3740 MB. Is this because of the RAM allocated to graphics? iPhone 11 would need the least RAM, iPhone 11 Pro would be in the middle, and iPhone 11 Pro Max would need the most.
2. Most of the entries are 2.66 GHz. However, two entries are for 1.86 GHz. I did some quick calculations and I'm wondering if the actual clock speed is not 2.65 GHz but in fact is 2.6666... GHz. If that 1.86 is actually 1.8666..., then that is EXACTLY 70% of full clockspeed. So I wonder if it's low power mode, which in this case cuts 30% off the clock speed.
3. At 1.86 GHz, the score is 783 / 1560 for Geekbench 5. That 30% number I quoted above doesn't make sense here in terms of performance, since that's only about half of expected performance. What could account for this? Is there a different mix of performance vs. efficiency cores at play here? Note that for 1.86 GHz, it still lists 1 processor with 6 cores like it does at 2.66 GHz, but at 1.86 Hz, the multi-core vs. single-core multiplier is 2:1. At 2.66 GHz, the multiplier is more like 2.6 or so.
4. There is indeed much variation in scores. Some of the worst scores include both the 11 and the Pros, and some of the best scores include both the 11 and the Pros. So, performance is not linked to model. The comparatively crappy scores we saw early on for iPhone 11 were most likely just because of variation in this benchmark, +/- throttling.
5. Most of the variation in scores has to do with multi-core. Single-core performance is mostly consistent, but the multi-core ranges much more from about 2200 to about 3550. I wonder if throttling is involved here, since presumably throttling would affect multi-core performance more than single-core, because of the higher heat generated from multi-core.