Our newest revisit could also be considered our oldest: the Nehalem microarchitecture is nearly ten years old now, having launched in November 2008 after an initial showing at Intel’s 2007 Developer Forum, and we’re back to revive our i7-930 in 2017.
The sample chosen for these tests is another from the GN personal stash, a well-traveled i7-930 originally from Steve’s own computer that saw service in some of our very first case reviews, but has been mostly relegated to the shelf o’ chips since 2013. The 930 was one of the later Nehalem CPUs, released in Q1 2010 for $294, exactly one year ahead of the advent of the still-popular Sandy Bridge architecture. That includes the release of the i7-2600K, which we’ve already revisited in detail.
Sandy Bridge was a huge step for Intel, but Nehalem processors were actually the first generation to be branded with the now-familiar i5 and i7 naming convention (no i3s, though). A couple features make these CPUs worth a look today: Hyperthreading was (re)introduced with i7 chips, meaning that even the oldest generation of i7s has 4C/8T, and overclocking could offer huge leaps in performance often limited by heat and safe voltages rather than software stability or artificial caps.
From a historical perspective, the gaming CPUs on our charts that competed with the $294 i7-930 were the $295 Phenom II X6 1090T and $199 1055T (launched later that year), as well as the $216 i5-2500K and the $317 i7-2600K (launched a full year later). That provides some context for how the 930 was holding up in 2010. When comparing these CPUs, it’s important to keep in mind that the Nehalem architecture had already existed for quite some time.
For a modern perspective: from a quick search on ebay, there are multiple listings for i7-930s at or under $40, compared to the overly optimistic $140-ish range that 2600Ks seem to be listed at, or even the $65-ish 1055Ts. Our revisits are typically approached from a “should I keep it” perspective, rather than “should I buy it,” but the i7-930 is selling for even less than G4560s--new or used--putting an interesting spin on these results.
CPU TESTING METHODOLOGY
GAME TEST METHODOLOGY
NVIDIA 376.33 drivers were used for benchmarking. Game settings were manually controlled for the DUT. All games were run at presets defined in their respective charts. All other game settings are defined in respective game benchmarks, which we publish separately from GPU and CPU reviews. Our test courses, in the event manual testing is executed, are also uploaded within that content. This allows others to replicate our results by studying our bench courses.
Windows 10-64 build 14393.1066 was used for testing.
Some benchmarks disable EIST, Turbo, and other features -- please check each section to learn if that is the case. Otherwise, for game benchmarks, assume stock settings (Turbo enabled). We always disable C-states.
Average FPS, 1% low, and 0.1% low times are measured. We do not report maximum or minimum FPS results as we consider these numbers to be pure outliers. Instead, we take an average of the lowest 1% of results (1% low) to show real-world, noticeable dips; we then take an average of the lowest 0.1% of results for severe spikes. GN originally coined the phrases “1% LOWs” and “0.1% LOWs” for these metrics.
EVGA Supernova 750 G2L 80+ Gold
HyperX Savage 32GB 2400MHz (4x8GB)
Corsair Force LE 240GB SSD
Open Air Test Bench
EVGA GeForce GTX 1080 FTW
Note: fan and pump settings are configured on a per-test basis.
970 (RD9x0) Platform:
-ASUS 970 PRO GAMING/AURA
-Phenom II X6 1055T (125W TDP)
-Phenom II X6 1090T
Core Components (Unchanging)
- NZXT 1200W Hale90v2
- For DDR4 platforms: Corsair Vengeance LPX 32GB 3200MHz
- For DDR3 platforms: HyperX Savage 32GB 2400MHz (note: only 2133MHz was supported on our SNB platform)
- Intel 730 480GB SSD
- Open Air Test Bench
- Cooler #1 (Air): Be Quiet! Dark Rock 3
- Cooler #2 (Cheap liquid): Asetek 570LC w/ Gentle Typhoon fan (this is the one we used for this particular article)
- Cooler #3 (High-end): Kraken X62
- Video Card: EVGA GTX 1080 FTW1
- Note: fan and pump settings are configured on a per-test basis.
- MSI Gaming Pro Carbon
- i7-7700K (x2) samples from motherboard vendors
- i5-7600K purchased by GN
- MSI Gaming M7
- i7-6700K retail
- Gigabyte Z97X G1 WIFI-BK
- MSI GD65 Z77
- i5 3570K
- ASUS Crosshair VI
Dx12 games are benchmarked using PresentMon onPresent, with further data analysis from GN-made tools.
- ASUS Deluxe Pro X58
Some Notes on Memory
Premiere & Blender tests do not exceed 8GB DRAM. Capacity is a non-issue for our testing, so long as it is >16GB. Memory frequency was set to 1866MHz for stock tests and run in triple channel mode (8GB x 3); the same timings etc. were used for OC tests, but with a higher frequency as noted in the next section. Functional memory settings were found by loading an XMP and tuning the frequency from there. To quote AnandTech’s 2008 preview, “Nehalem can get data out of memory quick like bunnies,” but this isn’t something that seriously affects our modern CPU benchmarks.
As usual, we’re taking a fairly casual approach to overclocking: our goal is only to do what a typical user could safely do with an hour or two of experimentation, not to break any records. Not every chip is created equal, and we don’t claim that all 930s will run at 4GHz, nor are we responsible for any damage resulting from overclocking.
We followed this popular guide from 2009, originally written for 920s. Overclocking the 930 was a more difficult process than our other revisits, as we could easily approach potentially dangerous temperatures (with Prime95) and voltages while still maintaining software stability. We made the decision to stop our experimentation at 4GHz, which was the target frequency of the guide linked above and kept temperatures under the rated safe temperature of 67.9C. Yes, we know it could go higher. We mostly used the “quick and dirty method,” although the chip we used was a D0 stepping variant, which potentially can run at lower voltages. Below are some of the settings used for our OC, which unusually didn’t cause any crashes during testing.
21 (* 191 = 4.011GHz)
10 (* 191 = 1910MHz, actually 1914MHz)
QPI Link Data Rate